UNIVERSIDADE FEDERAL DE SANTA CATARINA PROGRAMA DE PÓS-GRADUAÇÃO EM INGLÊS: ESTUDOS LINGUÍSTICOS E LITERÁRIOS

Meggie Rosar Fornazari

LOCALIZATION PRACTICES IN TRADING CARD GAMES: MAGIC THE GATHERING FROM ENGLISH INTO PORTUGUESE

Dissertação submetida ao Programa de Pós-Graduação em Inglês: Estudos Linguísticos e Literários da Universidade Federal de Santa Catarina para a obtenção do Grau de Mestre em Inglês. Orientador: Prof. Dr. Lincoln Fernandes

Florianópolis

2014

v

Acknowledgements

I would like to thank some people who played important roles in the process of these two years of academic work, consciously or not. To professors: Celso Tumolo, for teaching me how to write altogether; Raquel D’Ely and Rosely Xavier, for teaching me how to teach and how to write about teaching; to Maria Lúcia Milléo and Rosvitha Blume for presenting me with a closer look on Ernest Hemingway and post- WWII literature. Without your inspiration in undergraduate classes, I would not have managed to join PPGI. To Marco Rocha, for showing me the importance of grammar in corpus-based research; to Viviane Heberle, for opening my eyes to multidisciplinary and multimodal analyses; and Cristiane Vidal, for validating games research as something feasible and academic. To my advisor, professor Lincoln Fernandes, for being the devil’s advocate during revision stages, helping me to think critically about my ideas and teaching me about life in a university department. To my colleagues at TraCor, Grupo de Tradução e Corpora, especially Carlos Eduardo da Silva, for teaching me about working in cooperation with the research of others and creating a functional and efficient analysis system for us all to use. To my family, for the never-ending and ever-increasing support and pride in my academic life, work, travel, and endeavors. Especially to my academic mentors, Luciana and Jó, who have demonstrated to me that the road is long, but rewarding. To Volnei, for bringing me back to the game after 10 years, for being more proud of me than I was at times, for keeping my spirits up when I faltered, and for being at my side. To CNPq, for the financial support and for allowing me to teach English for 12h/week during my studies – a permission of great importance for my housing and sustenance in a city with expensive life costs such as Florianópolis. To Wizards of the Coast, owner of all rights pertaining to Magic The Gathering − a game which I do not own, but love to study. To people who are or have been part of their group and have helped me in the course of these two years: Mark Rosewater, Jennifer Meyen Robles, Gerald Blomgren, Aaron Forsythe, and Gavin Verhey of Wizards, professional players Paulo Vitor Damo da Rosa, Melissa DeTora, Michelle Olsen, and Miguel Gatica, Thaddeus Blanchette and Henrique Amigo of Devir, as well vi as Esther Jansen and Maryam Jaddoo of Alpha CRC − you have helped me to build legitimacy for my work. Finally, to the Magic player community in Florianópolis, especially to hobby stores Dragon's House and A Toca Revistaria − friendly and inviting environments to make friends and make Magic, as well as contacts and gather more ideas. To our local judges and long time players, for their timely aid in relation to the history of Magic when my short player history faltered. For all our Friday Night and Prerelease events, for the fun and camaraderie towards a female player that unfortunately is still not seen in other parts of the world. Also, to my Tumblr followers who have been close and encouraging watchers of my writing process. Spellslingers, this is only the beginning. vii

Abstract

The localization of a Trading Card Game of long durability and frequent updating such as Magic the Gathering (USA, 2003-present) requires a specific and detailed translation in relation to consistency, while following certain policies. This study presents a parallel, bilingual corpus created to observe technical constraints and localization policies which may pertain to the translation of such type of material from English into Brazilian Portuguese. In order to do so, the textual spoiler of 249 cards in each language has been extracted, aligned, and analyzed using Wordsmith 3.0 and Notepad++, with theoretical support of the translation description model created by Lambert & Van Gorp (1985) adapted for this genre. Macro analyses and subsequent micro analyses have both resulted in the observation of a translation product in Portuguese that is linguistically adequate to the target system, but not acceptable in the target cultural system. Additionally, certain sections of the product which require less rigidity in relation to the game rules could gain from the use of transcreation.

Keywords: Translation Practices, Localization, Trading Card Games.

viii

Resumo

A localização de um Jogo de Estampas Ilustradas de longa duração e atualização frequente como Magic The Gathering (EUA, 2003-atual) requer tradução específica e detalhada em relação a consistência lexical, ao mesmo tempo em que certas políticas devem ser seguidas. Este estudo apresenta um corpus paralelo e bilíngue criado a fim de observar restrições técnicas e políticas de localização que possam ser pertinentes à tradução de tal material do inglês ao Português do Brasil. Para tal fim, o texto de 249 cards em cada língua foi extraído, alinhado e analisado utilizando os programas Wordsmith 3.0 e Notepad++, com aporte teórico do modelo descritivo de traduções criado por Lambert e Van Gorp (1985) adaptado para o gênero. Ambas macro e micro análises resultaram na observação de um produto tradutório linguisticamente adequado ao sistema-alvo do Português, mas que não é aceitável em seu sistema cultural. Além disso, certas seções do produto que requerem menos rigidez em relação às regras do jogo poderiam ganhar em qualidade com o uso da transcriação.

Palavras-Chave: Práticas Tradutórias, Localização, Jogos de Estampas Ilustradas. ix

List of Tables

Table 1: Descriptive translation analysis model, adapted from Lambert & Van Gorp (1985) for the context of Trading Card Games...... 47 Table 2: Corpus details (Fernandes, 2004, p. 93)...... 52 Table 3: text selection criteria...... 53 Table 4: Preliminary data on the corpus metatext (Basic Rulebook in English and Portuguese) ...... 60 Table 5: Most frequent non-grammatical words (over 100 occurrences each) in each language...... 60 Table 6: Most common collocates with the word can (collocational horizon 1L and 1R)...... 61 Table 7: Semantic sorting of the 50 most frequent words in the studied corpus, as defined by the author ...... 68 Table 8: 10 sample card names from the studied corpus which exemplify the sophisticated vocabulary used in both languages, implying the use of such vocabulary as one localization practice...... 69 Table 9: 10 sample mechanics keywords from the studied corpus, accompanied by their reminder text...... 70 x

List of Figures

Figure 1: A Land-type card explaining a game mechanic (Exalted) returning in the Magic 2013 Core Set, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size)...... 17 Figure 2: Creature, Artifact, Sorcery, Instant, Planeswalker, and Enchantment card samples from the studied corpus. Source: Gatherer.wizards.com – 60% of actual size ...... 19 Figure 3: Plains, Forest, Mountain, Swamp, Island, and non-basic Land card samples from the studied corpus. Source: Gatherer.wizards.com – 60% of actual size...... 20 Figure 4: Creature-type cards with examples of static, triggered, and activated abilities ordered respectively from left to right. Source: Gatherer.wizards.com – 60% of actual size...... 22 Figure 5: The organization of the present study within the 12 areas of Translation Studies defined by Williams and Chesterman (2002)...... 23 Figure 6: A Goblin-supertype Creature card present in the studied corpus and its localization into Brazilian Portuguese. (Source: Gatherer.wizards.com – actual size) ...... 29 Figure 7: A Creature-type card and its localization to Brazilian Portuguese, which presents smaller font sizes. (Source: Gatherer.wizards.com – actual size) ...... 32 Figure 8: A Sorcery-type card, named Mind Rot, and its localization to Brazilian Portuguese, named Estiolar a Mente. (Source: Gatherer.wizards.com – actual size) ...... 34 Figure 9: A Creature-type card named Elderscale Wurm and its localization to Brazilian Portuguese, named Vorme Vetusto. (Source: Gatherer.wizards.com – actual size) ...... 36 Figure 10: Sample booster packs in both English and Portuguese. Sources: wizards.com and devir.com.br ...... 56 Figure 11: Credits for the Brazilian-localized Magic 2013 Basic Rulebook (extract of page 35). Source: wizards.com ...... 58 Figure 12: A Creature-type card named Aven Squire and its localization to Brazilian Portuguese, which presents reminder text (in italics and in parenthesis) about a keyword mechanic named Exalted. (Source: Gatherer.wizards.com – actual size) ...... 59 Figure 13: Parts of a Magic the Gathering card, as explained in the M13 Basic Rulebook (extract of page 5). Source: wizards.com ...... 64 xi

Figure 14: Word cloud with the 50 most frequent words in the analyzed corpus and their respective frequencies. Grammatical words of Brazilian Portuguese language have been ignored. Created at tagcrowd.com ...... 67 Figure 15: An Instant-type card which grants a static ability (First Strike) temporarily, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size)...... 72 Figure 16: An Instant-type card which flavor text references a minor character from another card named War Priest of Thune, and its localization to Portuguese (Source: Gatherer.wizards.com – actual size)...... 73 Figure 17: An Instant-type card which flavor text makes a real-world literary reference to Hamlet, by William Shakespeare, and its localization to Portuguese (Source: Gatherer.wizards.com – actual size)...... 74 Figure 18: Parts and steps of a turn during a match, extracted from the M13 Basic Rulebook, and its localization to Portuguese (Source: wizards.com). .... 76 Figure 19: A Sorcery-type card in which the flavor text makes a real-world cultural reference to the American expression it’s like Christmas in July, and its adequate localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size)...... 78 Figure 20: A Creature-type card in which both card name and flavor text make an in-game reference to another card in the corpus named Odric, Master Tactician, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size)...... 79

xii

Table of Contents

1 INTRODUCTION...... xiv 1.1 Context of Investigation ...... 16 1.2 A Research Object in Magic: The Gathering ...... 18 1.3 Locus Eloquentia ...... 23 1.4 Objectives and research questions ...... 24 1.5 Significance of the research ...... 25 2 REVIEW OF LITERATURE...... 27 2.1 Definition of fundamental notions ...... 28 2.1.1 Localization ...... 28 2.1.2 Internationalization ...... 32 2.1.3 Locale ...... 34 2.1.4 Translation Practice and Process ...... 35 2.1.5 Translation Policies ...... 36 2.2 Theoretical Foundations ...... 37 2.2.1 Some Studies on Localization ...... 37 2.2.2 Previous research on Localization of Games ...... 38 2.2.3 Research on Localization of Trading Card Games ...... 41 2.2.4 Corpus-based Research ...... 42 3 METHOD...... 45 3.1 Analysis Model ...... 46 3.2 Corpus tools ...... 48 3.3 Analytical Procedures ...... 49 4 A TRADING CARD GAME ANALYZED...... 51 4.1 Preliminary Data Analysis ...... 54 4.1.1 Presence or absence of genre, authors, localizers ...... 55 4.1.2 Metatext (Magic 2013 Basic Rulebook) ...... 59 4.2 Macrostructure Analysis ...... 63 4.2.1 Text division (card name, type, rules, reminder, flavor text) 63 4.2.2 Card name, type, rules, reminder text, and flavor presentation ...... 64 4.2.3 Internal narrative structure (reflection of plot onto the cards), dramatic plot (prologue, exposition, climax, conclusion, epilogue – reflected in the cards) ...... 65 4.2.4 Author comments (Framing Information) on rules description ...... 66 xiii

4.3 Microstructure Analysis ...... 66 4.3.1 Lexical choices ...... 67 4.3.2 Dominant grammatical patterns ...... 69 4.3.3 Speech reproduction (direct versus indirect) ...... 72 4.3.4 Narrative, perspective, and point of view ...... 73 4.3.5 Modality (passive versus active, uncertainty, ambiguity) .... 74 4.3.6 Language levels (archaic, popular, or dialect; game jargon) 75 4.4 Systemic Context Analysis ...... 76 4.4.1 Macro versus microstructure; text versus theory (norms or models) ...... 77 4.4.2 Intertextual relations ...... 78 4.5 Technical constraints ...... 79 4.6 Translation Policies ...... 80 5 CONCLUSION...... 83 5.1 Summary of the main points ...... 84 5.2 Limitations of the study ...... 84 5.3 Suggestions for future research ...... 85 References...... 87 xiv

15

1 INTRODUCTION

16

This chapter will briefly outline the context of the present academic investigation. The basic terminology has been defined under its selected scope and focus, as well as a short review of the literature related to localization of games. In sequence, the research objectives and guiding questions will be detailed, followed by the relevance for such a study to be developed.

1.1 Context of Investigation Translating is an activity that can be performed either individually or in groups, both under certain presuppositions of “choices, alternatives, decisions, strategies, aims, and goals” (Hermans, 1996, p. 2) which can be seen as unspoken policies. Such policies are usually defined by the client and abided by translators in order to produce a fitting result. When performed in groups which are geographically apart, technologies can now shorten distances, for example, through video conference calls – but nevertheless several decisions have to be made, informed, and justified together in order to ensure terminological consistency in the final product, as well as its quality1, while abiding by the recommended client policies. These group decisions are pivotal to change the translation product one way or another. According to Lambert and Van Gorp (1985) the translated product can be seen as adequate (following the target language rules and creating understandable text in that language) or acceptable (going beyond language-bound obligations in order to create a product that is also acceptable as something created specifically for the target audience).2 In addition, each translation service involves a certain amount of constraints that will limit the final translation product to some extent. In the case of games, translation is not the wording of choice; instead, localization is the term used. Localization as a construct will be more clearly defined in the Review of Literature of this study; however, in an initial stage it is possible to say that localization is the linguistic adaptation of a given product, similarly to translation, but cultural and technical aspects must also be taken into consideration (Palumbo, 2009 - see section 2.1 Definition of terms). In the culture of the early 21st century,

1 In the present study, quality will not be related to deeming a product good or bad, but instead as a result of the aforementioned secured consistency among cards. 2 The dichotomy between acceptable and adequate translations will be later defined in more detail. 17 the release of games is usually simultaneous worldwide, regardless of which media the game is released in (e.g., console games such as Mortal Kombat, board games such as Monopoly, card games such as Magic: the Gathering, social games such as Candy Crush, among many others). It is interesting to observe which decisions are made when those games are translated in a group, how the decisions are made, and more importantly, how they relate to client policies. A certain level of consistency is expected in the finished product, and all translators involved in the service should be supportive of the same translational choices being used on all instances where each is contextually required.

Figure 1: A Land-type card explaining a game mechanic (Exalted) returning in the Magic 2013 Core Set, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size).

Trading Card Games (TCGs), although having seemingly limited text, require specific care within the translation service, in order to reflect the game rules which players follow when playing the game itself (see Figure 1 above). However, there may be certain policies that are determined by other sections of the game development, such as Marketing, Copyright, or Research and Development, and must be followed under all 18 circumstances. For example, some terms may not be translated and maintained in English due to marketing orders, to sustain copyright and prevent legal complications. A more detailed definition of translation policies can be seen in the Review of Literature of this study. In this sense, this research focuses on the observation of the translation practices used by a localization group in the translation of a Trading Card Game based on the observation of a corpus with their translations, with a view to translation policies under which they must perform.

1.2 A Research Object in Magic: The Gathering Magic: The Gathering (MTG) is the name of the first Trading Card Game (TCG) to be ever produced, according to the Basic Rulebook. In TCGs, players may collect cards for displaying and trading, selling, or auctioning - but they may also use the cards to duel with 2 or more fellow players, building decks from cards of their possession, under several deck- building rules and game formats. Magic is representative of its genre because it is the first of its kind, and the biggest source to this day. 19

Figure 2: Creature, Artifact, Sorcery, Instant, Planeswalker, and Enchantment card samples from the studied corpus. Source: Gatherer.wizards.com – 60% of actual size

The game background story, which has been changing organically with the yearly release of new card sets and expansions, circles around the idea of the player acting as a planeswalker - a sort of wizard who has the ability of summoning Creatures, use Artifacts, and cast Sorceries or Instant spells (see Figure 2 above). Such ability is enacted when playing cards of such types, by paying a mana cost generated by Land cards (namely, the energy which fuels magic and is drawn from land – see Figure 3 below). There are five different land types which generate five different mana colors, which are generally guided by values of, for example, order and 20 protection for white, decay and death for black, deceit and intellect for blue, life and nature for green, and fury and chaos for red. Magic has the element of surprise within its game rules; as defined by Schell (2008), its rules may surprise one's opponent, and also the players themselves – in each and every card.

Figure 3: Plains, Forest, Mountain, Swamp, Island, and non-basic Land card samples from the studied corpus. Source: Gatherer.wizards.com – 60% of actual size.

The advent of the Internet has brought to light a need for consumers to have "product information, software, user manuals, games, educational materials" (Hartley, 2009, p. 106) available in their native language, which has sparked another need for localization services. Magic: The Gathering is currently localized into 10 languages, releasing approximately 1,500 cards 21 per year, each set and expansion being released simultaneously in all countries. The game is fully localized, in all its versions and aspects: "text, voiceover, manual, and packaging" (Chandler and Deming, 2012, p. 10). In the case of Magic, voiceover pertains only to the Core Set trailers released on YouTube every year. The languages to which cards are currently translated to are: French, German, Italian, Spanish, Russian, Simplified and Traditional Chinese, Japanese, Korean, and Brazilian Portuguese. For this reason, this study considers English to be the game's development language (Chandler and Deming, 2012), as this is the language in which the game is developed, and the source language to all of its localizations. It is also important to note that only one translation to Portuguese language is sold and used in all Portuguese-speaking countries in the world, although it is claimed to be for the Brazilian Portuguese locale. This implies, as expected for population reasons, that the majority of Portuguese-speaking players are Brazilian, and it reinforces the strength of Brazil as a locale as it is the only language on the MTG list that has a stated locale. The translational process may undergo several constraints, such as the number of characters available in the card and font size limits. Additionally, since every year a new Core Set and a narrative block of approximately three Expansions are released, there must be a consistency among all game characteristics such as its abilities, races3, back story, and rules. One point of interest in Magic the Gathering as a TCG is its use of keyword abilities. According to the Basic Rulebook, many cards have abilities that change the rules within a match. There are three types of abilities in Magic: static, triggered, and activated abilities. Static abilities are functioning in play at all times when a card is in play. Triggered abilities function automatically when a specific event occurs in play, for example, when a creature comes into play you draw a card. Activated abilities depend of specific actions enacted by the player, for example, sacrificing a creature in order to activate the ability. Examples of each of these abilities are in Figure 4 below.

3 In the case of fantasy games, it is common to use the word "races" instead of "ethnicity" because game races are usually mythological (e.g., goblins, angels, demons, homunculus, pegasi, avians, wurms, etc.) 22

Figure 4: Creature-type cards with examples of static, triggered, and activated abilities ordered respectively from left to right. Source: Gatherer.wizards.com – 60% of actual size.

As seen with the ability Lifelink on Figure 4 above, some of these abilities can be summarized in one keyword. This is rather common in Magic due to the frequent repetition of some of these abilities, and they are transformed in keywords in order to facilitate play. These two particular features of Trading Card Games will be of specific interest in Chapter 4.

23

1.3 Locus Eloquentia

Figure 5: The organization of the present study within the 12 areas of Translation Studies defined by Williams and Chesterman (2002).

According to Williams and Chesterman’s Translation Studies map (2002 - see Figure 5 above), the area in which this study is grounded on is that of Translation and Technology, more particularly in the subfield of Localization, as this is the very translational process to which the Magic: The Gathering (MTG) cards are submitted. This study is also based on the subfield of Text Analysis and Translation, more particularly in Comparison of Translations and their Source Texts. The text will be analyzed in terms of how it reflects the translation practices enacted by the team, as well as the translation policies established by their client. Therefore, the analysis will be performed both on word level and on sentence level. Some of the proposed objectives are to observe possible regularities of the translation team's behavior, and draw "general principles on how certain things get translated under certain conditions" (p. 7). 24

Still under Williams and Chesterman 2002, there is a field interface with Translation Process, in the subfield of Workplace studies, as the translation practices of the localization team will be analyzed; therefore, this proposal will take Translation Practices (post-service, procedural) into account, rather than (in-service, cognitive) Translation Process. While process implies studying the translators in-service “black box” and their cognitive actions (Williams & Chesterman, 2002), practice implies post- service observation of usual actions performed previous to or during the translation service. Magic: The Gathering is of particular interest in the sense that its localization is performed by a group of four people located in different parts of the globe, which implies a particular type of remote group decision-making process. This is the paramount focus of interest in this research. Section 2.1 provides further definition of the terminology used above.

1.4 Objectives and research questions This study intends to analyze the English into Portuguese localization of the 2013 Core Set of MTG cards, by studying the constraints that determine its localization practices, decision-making processes, and the policies which influence it. This piece of research is also expected to further develop the use of corpora for translation research. Some initial guiding principles which inform this study are that the European Portuguese linguistic system influences the Brazilian locale with some terminology, cultural references, or lexical choices that may seem excessively formal for the Brazilian linguistic system, therefore distancing such texts from the mainstream public. Another principle is that, due to a higher number of words necessary to convey the same rules, flavor text (i.e., text that has no effect on card play, but helps in setting the tone of a card into the Magic universe, according to the Basic Rulebook), and explanations to the players, font sizes are usually reduced in the Portuguese translation (see Section 3.3 for more information on flavor text). Stylistically, the MTG localization is expected to be directed to the source system, as linguistic and cultural terms are often extracted from the source system, while following linguistic expectations of Brazilian Portuguese – creating, therefore, an adequate translation product. The research questions related to this study are: • What are some of the main technical constraints involved in the localization of Magic: The Gathering cards from English to 25

Portuguese? (e.g., font sizes, terms or symbols which require no translation, consistency with previous editions, etc.) • What localization decisions could affect the final localized product of MTG? How? • What localization policies can affect their translation decisions? (e.g., copyright issues, marketing strategies, Research & Development style guidelines, etc.) The questions above exemplify the descriptive character of the present research, in the sense that this study will focus on the production of translated products, and not in the reception of this type of product.

1.5 Significance of the research The translation and localization of games is a field in constant growth; recently, there have been a few research studies of several levels on games localization performed at Universidade Federal de Santa Catarina. Each has different kinds of games as objects, namely an undergraduate final monograph on tabletop Role-Playing Games (RPG)4, a graduate study on video games and second language acquisition (Vidal, 2011), as well as a case study on computer-based Real-Time Strategy (RTS) and MMORPGs produced by Blizzard Entertainment5, among others. Trading Card Games have specific dynamics and constraints that are particularly interesting, as they rely on declarative speech during play, and each player has control of their own actions and consequences. MTG, in spite of having its mythology roughly based on RPG games such as Dungeons & Dragons, has pioneered in the game format and is now a well-established form of entertainment that is nonetheless worthy of study. Additionally, research on Trading Card Games translation is novel and has not been done either for the MTG linguistic pair to be observed in this study, or for any other – to the best of my knowledge. This study may open paths for the analysis of MTG localizations to other languages, and to the localization of other Trading Card Games as a game genre. Other more personal justifications involve the author’s personal experience with playing Magic The Gathering and wondering how its localization options have come to life after initial observations during her own participation in the player community. Having played the game

4 Coletti, 2013 and Coletti, 2014 (forthcoming). 5 Vidal and Elias, 2012 and Vidal, 2013. 26 casually during high school break and coming back ten years later due to the release of Magic computer games, a few terms such as Wall being localized as Barreira instead of Muro have brought my attention to how the language in the game is conveyed into Portuguese. After consultations to Portal Capes, BTS, and other academic data bases, and a rather fruitful course on Games with professors Viviane Heberle and Cristiane Vidal during the MA studies, the project has been consolidated into full-scale research.

In this chapter, the locus eloquentia of this proposal was set on the issue of analyzing the localization of games, localization team decisions, and the manner in which they may reflect given translation policies. The objectives, research questions, and guiding principles of the research were stated, being followed by the justification of the research. In the following chapter, the main constructs to be used in this research has been outlined and defined as well as the relevant literature on the issue. 27

2 REVIEW OF LITERATURE

28

This section of the study comprises key notions and concepts involved in the present research, followed by a review of the academic literature pertaining to the broad area of Localization zooming into Localization of Games. The Localization of Trading Card Games niche will also be defined as the territory of the present research.

2.1 Definition of fundamental notions In order to specify the theoretical affiliation of the present research, certain key concepts must be defined. With the following definitions in hand, a proper review of the specific literature can be read critically as they define the scope, notions, and concepts informing this study.

2.1.1 Localization One of the most relevant concepts for the present study is that of localization. Chandler (2012) limits the definition of localization to translating linguistic assets, namely "text and voiceover" (p. 4) without altering other game features, leaving that task to the internationalization processes to be defined in the following section. Since Magic is not often altered in other aspects, this definition also fits this study adequately. One exception that is beginning to fade into oblivion is an old Chinese- government taboo of not exposing bones, skeletons, and general images of broken limbs, which has caused selling prohibition of several games to the Chinese public in the past (Skoog, 2012), which can be clearly seen in gigantic internationalization services such as the covering of an entire skeleton army (Harper, 2007) in the Chinese expansion of the computer game World of Warcraft: Fall of the Lich King, produced by Blizzard Entertainment in 2009. In Magic, a small number of humanoid Skeleton creatures has been recently printed in Chinese expansions since 2010; initially, cards with non-human skeletal figures began getting printed, and after one or two years, less than 10 humanoid skeletal figures are now among the Chinese illustrations of MTG, according to their card game database, Gatherer (Gatherer.wizards.com).

29

Figure 6: A Goblin-supertype Creature card present in the studied corpus and its localization into Brazilian Portuguese. (Source: Gatherer.wizards.com – actual size)

According to Palumbo (2009), to adapt a product, in terms of the linguistic options in such product, would ultimately mean to translate each and every text element in it. When observing MTG it is possible to see that several terms remain without translation, such as the creature type "goblin", for example. This may be related to the fact that goblins refer to foreign cultures other than that of Brazil (see Figure 6 above)6; these issues have been more accurately addressed in the corpus micro analysis. However, the author also claims that such translation also involves "cultural adaptation, as target texts are required to reflect conventions and situations associated with the target market" (Palumbo, 2009, p. 71). It could be believed that not only conventions and standards can be observed, but also cultural adaptation in a broader sense – that of creating translation practices that sustain the public within comprehensible concepts, for example, with the visual aid of illustrations for players to see a goblin as the

6 Goblins are part of the mythos in several European civilizations, such as the Celtics. They are often confused with other mythological creatures from European folklore, such as trolls, elves, and dwarves (Bulfinch, 2009). 30

Magic mythology sees it. For a TCG, that would also imply to make rules understandable to the player simply by reading the card, and not having to resort to a reference book or a judge. Magic is marketed as a game you learn through playing (according to the Wizards of the Coast official website), and the Rulebooks are rarely consulted by players; MTG licensed judges usually study the Magic Comprehensive Rulebook more thoroughly as the judges themselves are the usual form of consultation during official events and tournaments.7 As localization services gather a wide volume of text to be translated, some information is often repeated. In addition, updates tend to recycle text. Therefore, such services are usually performed not by a single translator, but by a team of not only translators, but also assistants, revisers, technical text editors and other professionals, in order to fit translated text into the interface (Palumbo, 2009). Translation in the age of information should be seen as teamwork, including "project manager, terminologist, translator, reviser, DTP specialist, software engineer" (Hartley, 2009, p. 106) scattered throughout the globe, working online via remote servers. Additionally, for the aforementioned reasons, there is great tendency in utilizing translation memories (TMs) and/or term bases (terminology management systems). Therefore, localization is a modality in professional translation that resembles a production line, having improved the translational service as a whole in terms of "productivity, team-work, coordination and linguistic consistency" (Palumbo, 2009, p. 72). Localizing products or games is of commercial interest because simply selling a product abroad does not guarantee its sales, and Chandler (2012) claims that it is more likely for gamers to buy a game if it is localized into their mother tongue. This can be seen clearly with the increase in sales of World of Warcraft after its Brazilian localization release (Blizzard Entertainment, 2011). Chandler (2012) also mentions statistical data of approximately 30-70% of games sales being done outside its development country, and that range comes closer to 70% when players can enjoy their game in their home language. The concept of localization, therefore, is borrowed from Palumbo (2009), being the linguistic, cultural and technical adaptation of a given product into a specific market. Although such definition was coined by a

7 According to personal communication with Brazilian certifier judge (DCI Level 2) Henrique Amigo. 31 translation scholar with little specialization in the field of localization, it has been selected as the main definition for this study as it comprises all the desired individual features of localization that have been outlined by localization specialists, as seen above. Moreover, since Magic and Trading Card Games are a form of entertainment, localization is to be also understood here as Bernal-Merino (2006) describes the existence of one more layer of meaning, a creative layer that is set to "facilitate gamers' immersion in order to enhance the player's experience" (p. 6), which can be seen in Magic through flavor text, that is, illustrative text within the cards that add to the narrative text of the game. The volume of translatable text in MTG is surprisingly large, considering its small-sized medium, but it comprises all sources mentioned by Palumbo (ibid.): websites, user interface (within the cards themselves), product documentation, and 'collateral' materials (such as product packages). Therefore, in the proposed research, the terms translation and localization will be used interchangeably, both meaning localization itself as defined above.

32

2.1.2 Internationalization

Figure 7: A Creature-type card and its localization to Brazilian Portuguese, which presents smaller font sizes. (Source: Gatherer.wizards.com – actual size)

Chandler (2012) posits that internationalization is part of the product creation process, in order to create a product that will be easily localized into several languages with different accentuation, text direction, characters, and number treatments (in date, time, and currency) in several technical aspects: "user interface (UI), control scheme, game content", (Chandler, 2012, p. 4) among many others. Although Magic accepts graphic accents and Unicode characters (which include not only Latin-based characters, but also Asian characters and others), text direction is not changed for languages that are printed to be read from the right to the left, as it is in Japanese. Chandler also stresses the fact that, if pre-production internationalization is performed with attention to the international audiences to which the game is developed, the localization process should go about without issues and ruptures. Finally, she claims that the main objective of internationalizing a product is to produce a final localized version that feels specifically made for the locale, resulting in a game experience that is as similar as possible to the development language version (see Figure 7 above). 33

On the other hand, Hartley (2009) views internationalization as the set of adaptations necessary to fit several linguistic or cultural elements into the translation without the need to change design. It is important to note, however, that internationalization involves mainly the adaptations related to cultural expectations of a game, such as legal and numeric systems, age rating, and cultural adaptations related to the depiction of violence, which is even forbidden in some countries. One example of general internationalization would be the menu in global restaurant franchises such as McDonald's; the banana pop-tart is part of the menu in Brazilian restaurants only, not being served in restaurants of any other country (Nasr, 2006). When considering the development service in its entirety, it is of paramount importance to note that internationalization is a stage previous to localization (Chandler & Deming, 2012), as the cultural adaptations must be performed prior to the translation itself. This includes creating font and character formatting constraints such as using Unicode characters which allow for graphic marks and Asian characters, for example.

34

2.1.3 Locale

Figure 8: A Sorcery-type card, named Mind Rot, and its localization to Brazilian Portuguese, named Estiolar a Mente. (Source: Gatherer.wizards.com – actual size)

Another term to be defined is locale. Esselink (2000) points out that the Brazilian localization market is of medium range. The first group of languages to which localization is a priority is known as FIGS (French, Italian, German, Spanish), plus Japanese. In the second localization priority batch is Portuguese, Swedish, Norwegian, Danish, or Dutch. Out of these second-tier languages, only Brazilian Portuguese appears among the ten currently available localizations of Magic (see section 2.1 Research Object for the complete language list). A proper definition of locale would be a term used in localization to define not only the linguistic location of a target text, but also its geographic (and therefore cultural) location (Munday, 2009). Such concept is of particular interest to this research, since it is clearly stated that the Portuguese translation of MTG has Brazil as its locale, but it can be hypothesized that the Portugal locale may influence certain translational choices, such as card names with vocabulary that is rather rare in the Brazilian sphere (see Figure 8 above). 35

Internationalization and localization are two practices that are sometimes performed together. For Esselink (2000), an internationalized product should refrain from adding jargon and slang, as well as "culture- specific examples or references" (p. 25). This definition goes hand in hand with the expected behavior of the MTG localization, but it goes against the practice in, for example, games produced by Blizzard Entertainment such as World of Warcraft, which has quest names rather culturally specific, with Brazilian pop-culture references such as Silvio Santos, which follows the idea from the English version to use pop references also, but using localized references. In other words, these localized games have been under a process named culturalization, as just described. For this research, jargon created by the game (mechanic keywords and the like) will not be considered culture-specific jargon.

2.1.4 Translation Practice and Process Translation practice within the localization process is to be seen in this study according to the view of Lambert & Van Gorp (1985), as an open relation which depends both on the behavioral priorities set by the translators, and by the dominant conditioning factors of the target system. This is the idea of translation which permeates this study; however, it is only a part of the localization process. As mentioned before, every mention of process in this paper refers to the localization service and its usual procedures (see Section 1.2, Locus Eloquentia).

36

2.1.5 Translation Policies

Figure 9: A Creature-type card named Elderscale Wurm and its localization to Brazilian Portuguese, named Vorme Vetusto. (Source: Gatherer.wizards.com – actual size)

As mentioned earlier in the Introduction, translation policies are related to the implicit or explicit guidelines that limit a translation service in some aspects. One example pertaining to Trading Card Games could be names which cannot be translated because the Marketing or Legal departments would advise that copyright would be simpler to hold if the name were to be kept in English. Cronin (1998) claims that translation policies should be counteracted and/or manipulated by translators "for their own benefit" (p. 149). Munday (2009), on the other hand, defines translation policy in relation to which texts are selected for translation; such definition does not apply to the idea of translation policy in this study. Holmes (1972) says that a translation policy scholar should aid in defining "the place and role of translators" (p. 182). Therefore, it is a partial objective in this study to aid in defining the place and role of Trading Card Games localizers. Pym (1996, p. 170) mentions Brazil as a locale which typically brings the cultural content of media closer to home, therefore 37 domesticating translation (Venuti, 2012), in order to value fluency (Munday, 2009 – see Figure 9 above). The homogenizing convention by Sternberg et al. (1999) applies to MTG as it can be believed that the source text is seen as a so-called multilingual reality (that is, the in-game Multiverse, with its own terminology and mythos) which becomes believable when the reader-player willingly suspends disbelief (Baker, 2009). However, Toury (1995) adds to the definition the fact that policies also derive not only from the choice of texts, but also from people, namely publishing houses and the like, and the interface between text and publisher. This definition is more fitting to this study as it takes the publishing house into consideration, which in the present case would be the Wizards of the Coast Research & Development sector. It is important to note that scholars have outlined some translation strategies for localizers to use in order to bring to the reader-player an experience of the foreign culture within the target translation (Berman, 2004). This is a view that may be similar to the translation policy used in MTG, which may be confirmed or not in Data Analysis. To sum up, the view of translation policies that pertains to this study is a set of rules created by the development sector or publishing house, which must be followed by translators and localizers in order to yield a product that can be either domesticated to value fluency while sustaining cultural background of the target locale, or bring to the players a sample of the foreign culture by maintaining cultural information pertaining to the source locale − while at the same time sustaining the multilingual reality of the game story universe.

2.2 Theoretical Foundations Based on the basic concepts described above, this study will now gather history on the great field of Localization, narrowing it down to the Localization of Games area in order to arrive at the selected research territory for this study, which is the Localization of Trading Card Games.

2.2.1 Some Studies on Localization In addition to localization studies mentioned in the Introduction of this study, studies in the field of Localization have been directed more to software developers than to translators themselves, valuing pre-localization development (and internationalization thereof) much more than the 38 translational task at hand (Esselink, 2000; Chandler and Deming, 2012). Localization has recently received a more translational focus, however, in essential Translation Studies textbooks, which focus much more on the translational task and all it comprises than on how it should fit the industry (Munday, 2009; Palumbo, 2009).

2.2.2 Previous research on Localization of Games The industry has group endeavors such as the IGDA (International Game Developers Association) Best Practices in Game Localization (Honeywood, 2011). In this document, they bring forth an idea of culturalization developed by Edwards (2011), namely game localization that will allow more meaningful engagement on behalf of the players, ensuring that there is no offensive or contrasting content in the game which could hinder a more thorough engagement to the game itself. Also, four variables with potential cultural problems are listed (history, religion, ethnicity, and geopolitics). Finally, as the title presumes, some best practices based on experience of IGDA localizers are explained. Such practices are directed into publishing a game in a given locale in such a way that it would not suffer backlash or even governmental bans due to poor culturalization. Edwards adds that ways to do programming in order to create a localizable product are thoroughly described, and the localization section brings details of the service management for game developers, reinforcing the fact that Localization should be a part of the Research & Development process. There are also guidelines for proper localization project management. Interviews with localization companies can be seen in Lee (2013) in order to assess how the localization industry has been working with games, as well as ways in which this industry has dealt with the economic crisis started in 2008. Basically, games developers have decided for game dialogue subtitles instead of translation for dubbing and voice acting due to budget reasons. Additionally, in-house localization seems to be preferred over freelancers, and localization companies tend to branch out to other countries in order to explore appropriate taxes and offer better prices. Games tend to be smaller (in the case of independent or phone-based games), but the number of localization services has been increasing, and also the number of languages to which they are localized. He points out that Brazil is an emerging market for the games industry; that may be due to the fact that Brazilian people are very active in social networks, and also in the 39 games offered over Facebook, for example, or over telephones (up to 80% of the demand received by some localization service providers, according to his article). The same is true for "China and the Middle East" (Lee, 2013, p. 6), as well as more locales other than the usually initial FIGS block (French, Italian, German, and Spanish). This is true in the case of Magic in the sense that from the 10 localized languages, 4 are FIGS, 3 are in this emerging market, and there are also Russian, Japanese, and Korean versions. There is also mention of a preference by the industry to have remote localizers in order to have cultural understanding of each locale whenever possible. This is also true for Magic as half the localization team is Brazilian, living, gaming, playing Magic, and localizing in Brazil. Finally, case studies present technical information on some localization companies such as Testronic Labs (UK, 1998, Angry Birds and Batman: Arkham Asylum), and Localsoft (Spain, 1988, Pokémon Diamond and Wii Sports). Industry and the academia are brought together by Bernal-Merino (2006) with statistical data from the Entertainment Software Association in the USA, and bringing awareness to this US$7-billion media industry that increasingly requires games to be localized. Additionally, he outlines a brief history of electronic games from the pinball machine (1931) to Spacewar (1961) and the 1990 decade boom in video game popularity and acclaim. He also provides us with a definition of key terms pertaining to games localization, with a final direction to its translational, localizational, or transcreational aspects. The idea of transcreation as the actual practice performed by translators when localizing a game is presented by Mangiron & O'Hagan (2006; 2013, p. 196). Their definition derives from Mangiron's own experience in localizing games in the Japanese Final Fantasy series (used as a case study in their article). They claim that translators receive a rather high level of creative liberties to give the final localized product the intuitive feeling that it had been produced for the target audience. In the article, they bring forward an easy-to-follow comparison between software and games localization, as well as a comparison between screen translation and games localization. The previous is intuitively comparable, but screen translation only becomes comparable after 1999, with the 128-bit consoles such as the Sony Playstation 2, where DVD technology could hold more data such as subtitles and even dubbed voicing in more than two languages. Their main claim is that the final product should provide no oddities that 40 could disturb or truncate the game play or narrative experience, and that it is for that reason that localizers receive such creative liberties. The localization stage approaches games design and development in Donovan (2002), instead of remaining at the receiving end of the production line. Localizers responsible for very successful titles such as Diablo III, Assassin's Creed, and Fallout 3 provide game developers with 10 tips to make their product a very successful localization. Some of these include underlying values such as that the game style appropriateness should override grammar appropriateness, so much so that often beta test gamers are preferred instead of professional translators due to their greater immersion in the game context, situations, and flavor. Additionally, the culturalization element of localizing a product is paramount at Design stage, in order to prevent having to change a game dramatically later because of national issues on one locale (such as the aforementioned taboo related to exposure of bones in China). In turn, the player community is seen as fundamental in order to affect the industry (Djordevic, 2012), as well as to push such industry into localizing a game into their language. The example given is related to the Brazilian audience and game developer BioWare, producer of blockbusters such as the Mass Effect trilogy and Star Wars: Knights of the Old Republic. Another example given is that Sony usually produces games with 15 dubbed languages, and many more with subtitle localizations; Djordevic (2012) claims that this is due to the fact that the European audience expects to see games in their home language. Some countries are also mentioned as future market opportunities that, until the article had been released, had not been tended to. Finally, he stresses the importance of some extent of collaboration among localization specialists from different game developers or publishers, which can be seen in Game Developer Conferences. Localization and Law are connected (McCurley, 2011) as regards to creating proper justifications for localization choices, as well as to adapting the product or game into a given culture and its law system pertaining to games (McCurley, 2011). Regulation offices such as the ESRB (United States) have great impact on how a game is going to be localized. The example given is World of Warcraft (by Blizzard Entertainment) and its localization into Brazilian Portuguese. The legal connection comes with the use of precedents in order to make localization decisions: what to do, and what not to do, based on what has been done before, in a rather prescriptive manner. 41

The gap between the industry and the audience can be seen in Maragos (2009), who explains the localization process to the supporting gamers within the Atlus (game publisher of Tactics Ogre, among others) game forums. Industry and academia also come together in the work of researcher and localizator Coletti (2013), who announces corpus-based research on the Brazilian localization of three Dungeons & Dragons 4th Edition books. Another fruitful encounter is that of gamers and researchers. Vidal and Elias (2012) present their prized paper with a case study on Star Craft II, fully localized into Brazilian Portuguese, under the light of Newmark (1988). They have not only brought a short definition of terms and a sample of the work a localizator may usually receive, but they have also summarized Blizzard Entertainment's marketing campaign on the Brazilian public, as well as player reception seen in their community forums. Their main findings from the player community comprise the poor acceptance of the São Paulo regional accent for the voice acting, as well as the fact that players were not allowed to switch language options if they so desired, having only Brazilian Portuguese available. This last finding is particularly problematic as this stops people from using games as a language learning tool – a tool that has been used for generations to acquire not only English, but other languages as well. An interview by Cunningham (2013) of former Square Enix Japanese-English localizator Tom Slattery brings to light a few details on translator creativity previously mentioned by Mangiron & O'Hagan (2006). Creative freedoms given at Square Enix (developer of the Final Fantasy series, among others) is at carte-blanche level, but localization considers other limitations that are not part of the Square Enix policies, such as technical constraints, fan expectations, and time/schedule issues. This can also be seen in games produced by Blizzard Entertainment, such as Diablo, World of Warcraft, and StarCraft. In this interview, Slattery posits a similar point of view as Mangiron’s, saying that localized products "should be enjoyed as if it had been created domestically for the audience playing it" (p. 6), with the ultimate objective of making a fun game that "feels natural, immersive, and culturally appropriate" (Mangiron & O'Hagan, 2006, p. 6).

2.2.3 Research on Localization of Trading Card Games To the academia, MTG has been acquiring attention for other achievements: mostly, those related to game style patent (MacInnes, 2006); 42 rare and mythic-rare card auctioning, as well as marketing strategies (Walton, 1995); its impact on virtual cultures (Trammell, 2010); and as a reference to the popular culture started in the 1990 decade along with phenomena such as the Lord of the Rings, the Star Wars prequel trilogy, and the Harry Potter series (Whited, 2004). In the course of this two-year study, I (Fornazari, 2013) have published a paper on visual grammar and the synchrony between illustration, rules, and flavor text8 in Magic: The Gathering (popularly known as Magic). However, to my knowledge, there had not been any linguistic or translational research related to Magic until now. Although there have not been many academic papers related to the localization of Trading Card Games, there have been websites dedicated to informal commentary on the localization of Magic to Spanish (Grande, 2013) and Portuguese (Fornazari, 2013), particularly involving narrative expansions not pertaining to the Magic 2013 Core Set analyzed in this corpus. The present academic research has also been announced in the I Simpósio de Pesquisa em Games at UFSC in 2012, with a short paper published digitally (Fornazari, 2013). The aforementioned resources have been collected and obtained through extensive searches through Portal Capes, BTS, and Google Scholar.

2.2.4 Corpus-based Research In a corpus, one or more languages can be presented. Monolingual corpora such as COCA (Corpus of Contemporary American English), for example, which pertains only contemporary North-American English (En- US), cannot be parallel as they need two languages or more to be put in parallel with one another. A corpus can also be categorized according to the temporal restriction of its scope, being synchronic or diachronic. Baker (1995) describes corpora domains as being either general or specialized. An example of general corpora could be COCA, as it comprises sources from several different media and genres (e.g., television programs, magazines, among others). In the case of simultaneously-shipped (sim-ship) games, released simultaneously in several languages, one can consider the original

8 As defined by the Basic Rulebook, flavor text is part of the textual elements in a card (usually located on the bottom of the card text, in italics) with information on the fictional worlds in which the stories of the game unravel. It is important to note that this type of text has no function in game play. 43

L1 in being what Chandler and Deming (2012) call the game development language. More information on corpus details for this study is available in Chapter 4.

This chapter has presented the main notions pertaining to this study, followed by a brief review of the specialized literature in the area of Localization of Games. In sequence, the niche of Localization of Trading Card Games has been defined as the main territory of this research. In the following chapter, the method to be used in the proposed research, through the creation of a multimodal corpus and its analytical procedures is thoroughly explained. 44

45

3 METHOD

46

This chapter explains in detail the method applied in the present research: the creation of a bilingual corpus and its later data analysis, performed in order to observe the translational practices pertaining to this type of translation service.

3.1 Analysis Model The model used to analyze the corpus was created by Lambert & Van Gorp (1985), adapted for Trading Card Games. It is a broad model that is based on a general and flexible translation theory; therefore, there is room for adaptation. This choice has been made because Lambert & Van Gorp make explicit that their system is not necessarily only appropriate for literary systems. Based on their statement, it could be adapted for a Trading Card Game such as Magic. Through such a model, it will be possible to analyze the resulting text from the point of view of its locale. In their pivotal 1985 article, Lambert and Van Gorp have created a hypothetical model to describe translations that is broad enough to study them within a general and flexible theory of translation. Their objective was to observe relations which permeate the production and format of translations, through a functional and semiotic approach. Ultimately, they aim to characterize textual and translational strategies through that observation and description. Their analysis is conceptualized in two levels: macrostructure (more general) and microstructure (more specific); they are both interrelated in another stage of the analysis which they have named systemic context. However, their model was directed towards the analysis of literary translations; therefore, some adaptations to the genre of Trading Card Games were necessary. These adaptations pertain basically the object of analysis. When Lambert & Van Gorp (1985) observed in the macro level "chapter titles, act presentation, and scenes" (p. 211) I have changed that internal text division to that of a card, namely card name, type, rules, and flavor text. Therefore, the changes made are related not to the model structure itself, but to adapt the textual genre of trading cards to their analysis model (see Table 1 below). 47

Table 1: Descriptive translation analysis model, adapted from Lambert & Van Gorp (1985) for the context of Trading Card Games. 1. Initial data 1.1 Presence or absence of genre, authors, localizers 1.2 Metatext (M13 Rulebook) 2. Macrostructure 2.1 Text division (card name, type, rules, flavor text) 2.2 Card name, type, rule and flavor presentation 2.3 Internal narrative structure (reflection of plot onto the cards), dramatic plot (prologue, exposition, climax, conclusion, epilogue – reflected in the cards) 2.4 Author comments (Framing Information) on rules description 3. Microstructure 3.1 Lexical choices 3.2 Dominant grammatical patterns 3.3 Speech reproduction (direct versus indirect) 3.4 Narrative, perspective, and point of view 3.5 Modality (passive versus active, uncertainty, ambiguity) 3.6 Language levels (archaic, popular, or dialect; game jargon) 4. Systemic context 4.1 Macro versus microstructure; text versus theory (norms or models) 4.2 Intertextual relations

One aspect of the macro analysis to be observed is whether the localization is oriented predominantly towards the target system (acceptable) or to the source system (adequate). As defined by Lambert & Van Gorp (1985), such dichotomy between adequateness and acceptability is subjective and translations do not tend to be completely coherent in relation to either. At the same time, the categorization and judgment of an acceptable translation choice will be performed while considering in detail the source situation, instead of merely listing the differences between source and target texts (Lambert & Van Gorp, 1985). In accordance with the assumptions that guide corpus-based research in translation studies listed by Maeve Olohan (2004) this research is set to find probable and typical features of the MTG translation, as well as interpreting what can be seen as unusual in it. The analysis is set on similar objectives to those of Mason (2001), namely understanding the translation practice of a localization team as regards to the particular context of 48

Trading Card Games as a genre, with its particular "discourse, text, and rhetorical purpose" (Olohan, 2004, p. 30). The major focus of interest in this analysis is to perceive what Olohan (2004) defines as style-related subconscious behavior on behalf of the translators, in opposition to conscious alterations performed in order to meet game requirements, be it in relation to the game rules, copyright issues, style guides, among others (in other words, its translation policies). Additionally, the subconscious style of one translator may not reflect on the translations performed by another member of the translation team; therefore, there may have been some extent of group decision-making processes as regards to one or another translation choice that is related to the style of one member of the group. One of the main uses of this corpus in this research is to observe and understand stylistic conventions and usual translation strategies that can be seen as typical of the specific text genre (Olohan, 2004) that is a Trading Card Game.

3.2 Corpus tools Corpus tools and analyses have been used as part of the empirical method informing this research, being corpus as defined by Olohan (2004) as an electronic selection and compilation of texts, to be analyzed through specific corpora software. It is important to note that texts within such definition can be written, spoken, or multimodal. Corpus methods have been chosen due to several factors, from the multimodality seen in the cards to its research practicality when dealing with authentic material digitally. Another reason is its fruitful use in analyzing linguistic (and therefore translational) performance rather than competence. According to Leech (1992), corpus-based research enables us to visualize externalized language and therefore to analyze phenomena that can be observed through the finished product, outside the translator's mind, and made independent from it. This goes against Olohan’s (2004) claim that parallel corpora are used without much focus on the translation activity as a whole; however, this corpus analysis should provide insight on the aforementioned phenomena. In this research, the objective is to make observations and draw conclusions under that basis - the translators' mind, real-time processes during the translation service is not under the selected scope. The corpus assembled in this research is, at an early stage, static. That is, the Magic 2013 Core Set is a fixed collection of cards that depicts 49 the game at a specific point in its time, storyline, and history. The research also aims to become corpus-driven research, since there is no pre-existing theoretical description of the Trading Card Game genre. Therefore, the corpus created in this study is bilingual, parallel, and static, in order to perform corpus-driven research on the textual genre of TCG. Both English and Portuguese versions of all MTG cards are available online at its official database website, Gatherer, for no charge, in both JPG (image) and textual formats; this occurs because only original cards have marketing value and forging MTG cards was until recently considered to be nearly impossible (MacInnes, 2006). The 2013 Core Set consists of 249 cards and a bilingual corpus has been constructed by gathering the textual spoiler of both English and Portuguese image files, for a total of 498 cards. Additionally, its metatext (the Basic Rulebook) has also comprised the corpus. Text alignment was manually performed by utilizing the textual spoiler of the cards. The corpus will eventually be made available online within the COPA-TRAD9 system, currently under development at UFSC with data from a number of fantasy book series.

3.3 Analytical Procedures Although the proposed research follows and adapts Lambert & Van Gorp's model, the procedures are rather similar to what they had proposed in 1985. In a first moment, the macrostructure features of the translated cards have been collected and analyzed briefly, as regards to the game being identified as a translation, with translators and localizers given due credit, its recognition as a translated text in relation to "linguistic interference, neologisms, socio-cultural features" (Lambert & Van Gorp, 1985, p. 206), and its meta-text (i.e., the M13 Basic Rulebook, available gratis in the official Magic website in both languages, in PDF format). Such first-level analysis has provided insight on the translation practices adopted by the localization team, as well as a selection of part of the corpus for further micro analysis. Such micro analyses of the corpus have been performed based on the preliminary macro analysis. At this stage, statistical data has been gathered and analyzed in relation to the insertion, deletion, explicitation or alteration of information within the text. This second-level micro analysis has

9 Available at the Grupo de Tradução e Corpora website at www.tracor.ufsc.br 50 provided data on the consistency of such adding or deleting information, as well as given further and more precise visualization of the translation practices that permeate such localization service. When necessary, such consistencies have been categorized and related to one another. Bernal (2006) claims that such dichotomies between, e.g., adequacy and acceptability are inherent to polysystemic communication, and therefore, to translation itself. When quoting Umberto Eco (2003, p. 100), decision-making processes are strongly featured in translation when he says that "To choose a target- or source- oriented direction is, once again, a matter of negotiation to be decided at every sentence." In this research, since the localization service is performed in a group of four people, the focus of analysis is on observing which decisions seem to have been made in a group, individually, or even subconsciously− without the translator having consciously thought about it in practice. Corpus alignment was performed manually by segment, by utilizing parallel concordancing software. In sequence, concordancing queries were sorted out in different ways, in order to observe usage patterns. Searches for card data analysis were performed through bilingual concordancing software WordSmith 3.0 (Mike Scott, 1999), in order to assess specific terms and their equivalents. Frequency lists have been retrieved in order to check for ability keywords and their equivalents in Portuguese, particularly to assess consistency in translation. It is expected that the type-token ratio of the MTG corpus in English comes to a lower figure than in Portuguese, as the repetition of several card types, races, classes, and abilities is expected to be very frequent. Collocation has been observed, for example, as to which verbs are followed or preceded by card abilities. In order to do so, the collocation horizon will be set to one word to the left, and one to the right. Semantic prosody has been analyzed in terms of the positive or negative connotation given to certain collocates (e.g., for verbs following the term 'creature'). 51

4 A TRADING CARD GAME ANALYZED

52

This chapter presents quantitative data on the bilingual, parallel corpus constructed according to what is indicated in Chapter 3 (Method). Firstly, the corpus will be described according to its nature. Later, the adaptation of a model by Lambert & Van Gorp (1985) will be used to analyze the given corpus, from its macro and micro perspectives. The results here will also provide qualitative data on the translation practices pertaining to Trading Card Games, to be further discussed on Chapter 5.

As previously mentioned in the Method section of this study, a parallel and bilingual corpus has been built in order to analyze typicalities pertaining to the translation practices of this specific genre of localizing Trading Card Games. According to Fernandes (2004), such corpora are classified according to the parameters seen below on Table 4.

Table 2: Corpus details (Fernandes, 2004, p. 93). Languages American English (En-US) and Brazilian Portuguese (Pt-Br) Time Synchronic (MTG Core Set 2013) Domain Specialized (Trading Card Games genre) Directionality Unidirectional (En-US to Pt-Br)

The corpus created for the present study is bilingual; however, there could have been more languages. For example, Rocha (2013) presents a parallel corpus of the song Garota de Ipanema in six languages. This study has limited scope on one single Core Set of the game, being therefore synchronic as it analyzes the instances of the localized game at this particular point in time. The studied corpus is specialized in the sense that it is limited to Trading Card Games as a textual genre. Having the idea of development language in mind, the present corpus is unidirectional as it is developed in American English and is later translated and localized to the other 9 released languages. However, it is important to note that every set or expansion is released in all 10 languages on the same dates. In turn, corpora can be bidirectional when comprising originals and their translations in both languages. This could be done on the topic of Trading Card Games with its critique. Critique articles, written by specialized game websites and blogs, are often translated to English 53 although having been originally released in other languages. However, this is not under the scope of the present research.

Table 3: text selection criteria. Period Released in June 2012 Typicality Games company (Wizards of the Coast, under Hasbro) Mode and Digitalized version of cards printed to be played medium Overall 51,945 words corpus size

The Magic 2013 Core Set was released in June 2012, a moment which has proven to be practical for the data collection of this study, as it was the first yearly Core Set of Magic cards released during its research time span. The Core Set has been chosen due to the fact that Core Sets are more general releases that serve as a narrative re-stabilization from the previous narrative block, wrapping up loose ends of that storyline, and preparing the ground for the next narrative block to be released in a few months (see Section 4.2.3 for further narrative explanation). The typicality of this corpus is particular, due to the fact that it is a game – and games are usually translated either in-house or by third-party companies selected by the game developer. That is the current case of Magic, which is localized by a professional localization company situated in Barcelona, Spain. For many years, the game had been translated by Devir, the company which has been importing Magic to Brazil since the Fourth Edition until today (although they presently do not have exclusivity importing contracts). During Lorwyn, a whimsical narrative block released in 2007, the localization was moved out of Brazil, but Brazilian translators continued to be hired. These facts indicate that this corpus is restricted to one particular view – that of the game localizer, with the best interest of the game developer. In Brazil, the rating system indicates that Magic is for 13- year-olds or older, due to simulated violence and references to blood and gore. Although the game is commercialized in paper cards, and both languages are available in Brazil in paper format, the mode and medium of this corpus was selected as the transcribed text of digitalized cards, as they are easily organized and retrieved at the Magic the Gathering official 54 website in all 10 languages. It is important to note that this research is for purely academic purposes within the field of Translation Studies only, and the selling rights of MTG in all its forms (paper-based, online, and computer games) belong to Wizards of the Coast since 1993; the company was bought by Hasbro in 1999. Therefore, all rights on cards, characters and their likenesses, and mana symbols belong to Wizards of the Coast and Hasbro, which are used simply because their product, Magic The Gathering, is a pivotal example of Trading Card Games as a genre (Fornazari, forthcoming). According to Fernandes (2004), corpus sizes should fit their research objectives, in order to analyze phenomena in it. The M13 Core Set consists of 249 cards in each language, and its Basic Rulebook has been added to the corpus in the guise of metatext, totalizing 51,945 words. This corpus size is particularly small in comparison to others, but it is important to consider the media which is being analyzed. Cards are small and, considering its usual printing area, there is considerable amounts of text in a trading card game like Magic. Other Trading Card Games, such as Pokémon and Yu Gi Oh, do not present as much text on their cards. This is probably due to the fact that they do not carry any narrative flavor text. Another important fact to consider is that the metatext comprises 32,955 words (63.4% of the total corpus size), which could indicate low representativeness of this corpus. However, it is important to remember once more the size of the analyzed media and consider the fact that 498 cards have 18,540 words in total – with an average of approximately 37 words per card. It is interesting to note that for each language, the average number of words per card is discrepant: it is possible to calculate averages of 35 words per card in English in opposition to 40 words per card in Portuguese. These averages are consistent with the usually longer length of localized text (Chandler and Deming, 2012).

4.1 Preliminary Data Analysis Our systemic analysis begins by gathering first-look data from the corpus, in order to bring other considerations to surface later on, during its macro and micro analysis. According to Lambert and Van Gorp (1985), this initial information collection should provide an approximate overview into the probable general strategy for translation, as well as their main strategic priorities (Lambert and Van Gorp, 1985, p. 206).

55

4.1.1 Presence or absence of genre, authors, localizers Since the corpus is not compiled in a full book, this initial heading analysis and bibliographical information can be retrieved from some of its product packaging, which is fully localized. The genre of Trading Card Games is not clearly stated in this section, although it is practically the second title of their biggest sales product: the booster pack. Each booster pack is a blind bag with 15 semi-random cards inside. All booster packs carry 1 Land card, 10 Common cards, 3 Uncommon cards, and 1 card that may be either Rare or Mythic Rare. There may also be an extra foil card of any rarity, a token card (for temporary creatures), or advertisements about the present or future expansions and other Magic The Gathering products. Players buy booster packs in order to try their luck in opening cards that are sometimes more valuable than the price paid for the booster itself. In 2012, the Magic 2013 booster packs cost approximately R$11.50 in Brazil. Therefore, players technically pay R$0.75 per card; however, although many of the commons are worth less than that, the uncommon cards often surpass this value and usually the rare card makes it a fair trade. The trading value of cards is not only based on their rarity, but also on their potential during play on certain game modes, and particularly on how they are played in professional tournaments around the world. These tournaments and Grand Prix are broadcasted online on the Magic The Gathering official Youtube channel, making front page news at their official website, as well as being covered by many other competition gaming specialized websites. Some rare or mythic rare cards may be speculated to be worth R$40.00, R$70.00, or even R$120.00 at the moment of Prerelease events, only due to their potential in professional standard matches. Due to its importance as its main product, the booster pack is representative of the genre and could be seen as an analogue to a book cover (see Figure 10 below).

56

Figure 10: Sample booster packs in both English and Portuguese. Sources: wizards.com and devir.com.br

The Brazilian booster packs are explicitly stated as a translation by stating the target language at the upper left corner of the package. However, simply adding the language to the package gives room to consider if this is a “translation, an adaptation, or a form of imitation” (Lambert & Van Gorp, 1985, p. 206). As mentioned earlier, since the game is simultaneously released in all countries, the English version may not be seen as an original, but as a development language. In the Brazilian version, however, the genre of Trading Cards is slightly more explicit than in the development language, because it names the booster pack as a product by explicitating its content: 15 estampas ilustradas. Except for the name of the game, all other information is translated; however, not translating the name of the game and the translational choices used in the product packaging indicate that the product is translated with source-system orientation (which Lambert & Van Gorp, 57

1985, p. 202 conceptualize as an ‘adequate’ translation). It fits the grammar needs of Portuguese language, but it does not drift very far from the most literal translation options available – even if those translational choices may not be very intuitive. It is possible to assume that the term cartas has not been used due to its bad connotation in relation to casino games, such as Blackjack and Poker, or even divination (such as the reading of Tarot). When the game began being translated to Portuguese, circa 1997, probably the borrowing of the word card would not have been well accepted by the Brazilian public. However, in Brazil the word estampa is usually collocated with clothing, such as printed T-shirts (empirical data on such collocation can be verified with the use of a monolingual corpus search such as webcorp.co.uk). At the same time, the word ilustradas refers to visual aids in books, magazines, and the like. Since the illustration of Trading Cards has this in-game function of visual aid, the idea fits comfortably; a more fitting suggestion could be Cartas Ilustradas. The idea of illustrated cards could release the gambling connotation of the term cartas, while adding the narrative idea of visual aids which are rather common in illustrations within children’s literature (Fernandes, 2004). More preliminary data can be retrieved from the Credits section of our metatext, the Magic 2013 Basic Rulebook (see Figure 11 below). Although there is no clear reference to TCG as a genre, this can be implied from the copyright information on the bottom of Figure 10. Several developers are given credit to in different forms of authorship (game design, graphic design, rules editing, art direction, editing in general) and there is explicit mention and crediting of translation and localization as two separate functions. The two translators are Brazilian, one of them being a MTG judge and owner of a hobby store in Porto Alegre; this indicates that the translators are native speakers of Brazilian Portuguese and reside in the country, having contact with the country’s cultural expectations (Chandler and Deming, 2012). However, the in-house localizers are not Brazilian, which may imply their larger responsibility in managing the localization project and other activities which pertain to localization.

58

Figure 11: Credits for the Brazilian-localized Magic 2013 Basic Rulebook (extract of page 35). Source: wizards.com

It is important to note that, in fact, Richard Garfield is the original game designer of Magic, and not its original graphic designer as the Portuguese localized Rulebook states. This is probably a translation error, especially because the same credits are mentioned 8 lines below, then with the rightful original graphic designers. This error may have derived from improper recycling of translation memory units, and helps in evidencing that Computer-Assisted Translation (CAT) tools are used. There is metatextual information in the cards in relation to game mechanics , but they cannot be considered translation footnotes. Game mechanics would be keywords used in-game in order to mean a certain rule is a possibility during play. For example, in the mechanics called Trample, an attacking creature with power 6 that is blocked by a creature with toughness 4 is going to trample over the blocking creature, killing it on the way, and deal the additional 2 damage on to the defending player. This is not, however, a translational footnote, it is actually a game footnote that is also used in the development language. This is called reminder text, and it can be seen in italics and in parenthesis, next to the mechanics keyword. Figure 12 (see below) shows an example of reminder text that explains how the returning Magic 2013 mechanics, Exalted, works in play. 59

Figure 12: A Creature-type card named Aven Squire and its localization to Brazilian Portuguese, which presents reminder text (in italics and in parenthesis) about a keyword mechanic named Exalted. (Source: Gatherer.wizards.com – actual size)

4.1.2 Metatext (Magic 2013 Basic Rulebook) The choice of the Basic Rulebook as metatext was due to the fact that it was released at the same time as the Magic 2013 Core Set, and because it literally functions as its metatext, that is, a text that talks about the cards and how they work in play. Also, the Basic Rulebook is available gratis in PDF format at the MTG official website, being easily accessible by players of every level. The analyzed metatext is expected to enable the finding of features that are typical of translating this game genre, as well as the game itself, allowing for the observation of stylistic conventions and usual translational strategies which compose the genre (Olohan, 2004) of Trading Card Games. 60

Table 4: Preliminary data on the corpus metatext (Basic Rulebook in English and Portuguese) En-Rulebook Pt-Rulebook tokens 16.104 16.894 types 1.368 1.762 type-token ratio 30,53 32,54 sentences 1.025 745

The type-token ratio of the metatext was 30.53 in English and 32.54 in Portuguese (according to Table 4 above). These figures are quite close to expectation, although the text is relatively short – comprising of approximately 16,000 words in each language. Additionally, the higher TT ratio of the Brazilian cards indicates the higher lexical density pertaining to the language (Fernandes, 2004). It is also believed that the type-token ratio figures in this range are due to the repetition of keyword abilities and game rules, and because half the Rulebook is a game glossary.

Table 5: Most frequent non-grammatical words (over 100 occurrences each) in each language. En-Rulebook Pt-Rulebook Word Frequency Word Frequency you 425 criatura 294 creature 295 você 291 card 271 cards 263 ability 249 habilidade 259 can 190 mágica 180 spell 143 pode 148 player 133 jogador 135 game 130 mana 124 mana 125 dano 116 damage 113 jogo 116 -- -- terreno 103 -- -- tipo 101

In each text, frequency lists have been analyzed prior to alignment stage, and non-grammatical words with frequency over 100 were listed above on Table 5. In the case of nouns, each respective plural has been merged with its singular lemma. From this list, one word which calls more 61 attention is can (and its expected correspondence, pode), as it is the only verb with over 100 occurrences in each list. All possible conjugations of pode in Portuguese have been included as its lemmas, and can’t has been added as a lemma of can.

Table 6: Most common collocates with the word can (collocational horizon 1L and 1R). En-Rulebook Pt-Rulebook x can can x x pode pode x you (90) be (37) você (48) ser (29) player (12) cast (19) não (29) conjurar (12) creature (11) have (15) só (16) jogar (11) -- play (12) -- ter (10)

Departing from the aforementioned frequency lists, the most frequent co-occurrences (over 10 each) with collocational horizons immediately before and after can and pode have been obtained and listed (see Table 6 above). The analysis was limited to the first word next to can/pode as most occurrences were accounted for in this collocational horizon. Their concordance lines have been later exported from Wordsmith 3.0 to Notepad++ (Notepad++ Team, 2013) for manual alignment. After manual alignment, 161 segments10 have been selected based on their use patterns, in order to assess certain specific terms and correspondents, based on the 4 types of translational correspondence conceptualized by Martha Thunes (1998). These 4 correspondence types are literal, word-by-word translation (type 1), near-literal translation with grammatical insertions, deletions, or adaptations (type 2), syntactic but not semantic changes (type 3), and both semantic and syntactic differences in the target text (type 4). Segments have been selected as analysis unit, corresponding to short sentences, or parts of longer sentences cut automatically by Wordsmith in order to be only one line long. Amongst the 161 segments, there is segment intersection due to double concordances; for example, the expression you may cast corresponds both to concordance with you and with cast).

10 The notion of segment used for this study is any string of text separated between two final stops. As for the card textual spoiler used later, the full text of each card was considered in its entirety for alignment. 62

Most co-occurrences have been categorized as Type 2 in the Thunes framework, that is, nearly literal translation with differences in word order and use of grammatical words due to syntactic restrictions typical of Brazilian Portuguese language which DO not pertain to the English language. Many segments are 100% similar, but are located in different places within the Rulebook and are considered to be separate occurrences. It is possible to infer the use of Translation Memory (TM) systems in this translational service, in order to sustain consistency between each step and phase of the game, as explained in the Rulebook. In very few cases, some drastic changes are made in word order. This brief analysis provides insight on some discrepancies found in Table 8, where the frequency of several co-occurrences is quite reduced in Portuguese, in comparison with English. There are additions of só in several terms in which there is no correspondence in English, having 2 only can for 16 só pode. Altering the collocational horizon to only x can (2) and can x only (9), it is possible to find 11 more occurrences, but 3 remain with no correspondence found. For them, it can be assumed that só has been added as Framing Information, that is, information added by translators to clarify potentially difficult terms (Gile, 2009). In sum, it is possible to observe a translational pattern of a Type 2 majority. This is expected due to the metatext genre: Being a Rulebook, the Message to be conveyed (Gile, 2009) is of normative character, which is comparable in due proportion to legal text. The Basic Rulebook is a quick source for players to solve game doubts. However, in official sanctioned events the consultation of rules is done verbally by consulting Magic “judges” – volunteers who are tested and certified by Wizards of the Coast. These judges have ample knowledge not only of what pertains to the Basic Rulebook, but also of the Comprehensive Rules book, directed to judges only, which is also available gratis at the game official website. The absence of Thunes Types 3 and 4 in the metatext indicates that choices with different syntactic (Type 3) and/or semantic representations (Type 4) have not been privileged. This is another indication that the translation is produced to be adequate, according to Lambert & Van Gorp (1985): adequate to the linguistic demands only, but not concerned with public acceptance. It is also possible to hypothesize that these translational choices may not be encouraged in the briefing the localization team receives. Pragmatic equivalence seems to be privileged above all, perhaps partly due to the text genre, or due to the product localization briefing. One 63 of the objectives of analyzing the metatext is precisely this: to gather interesting patterns and issues on the corpus in order to later perform a closer micro analysis and draw observations on how this translation was performed.

4.2 Macrostructure Analysis According to Lambert & Van Gorp (1985), analyses performed on macro levels involve observing how the text is divided, titled, and presented; also, the types of narrative, dialogue, and description featured, as well as its internal narrative structure (plot, intrigue, and poetic structure), and authorial commentary. These observations should provide further data and bring to light hypotheses on microstructural strategies to be studied further in the next section.

4.2.1 Text division (card name, type, rules, reminder, flavor text) A fitting illustration of the card text internal division can be found in its metatext, the Magic 2013 Basic Rulebook (p. 5, as seen in Figure 13 below), featuring and explaining what text pertains to each part of a card. This structure has been changed to what it is presented today in the game’s Seventh Edition, which is the landmark between two game modes: Classic and Modern. The card design in essence (that is, the positioning of the parts below) has not been changed, but some design backgrounds have been made easier to read in comparison to the Classic cards. Since the back of the cards has never been changed, cards from all editions can be used in play modes such as Vintage and Legacy. 64

Figure 13: Parts of a Magic the Gathering card, as explained in the M13 Basic Rulebook (extract of page 5). Source: wizards.com

4.2.2 Card name, type, rules, reminder text, and flavor presentation It is important to note in Figure 13 above that the four sections named Mana Cost, Expansion Symbol, Power and Toughness, and Collector Number are not translated, being either numbers or non-verbal information. Illustrations are rarely localized, as mentioned in the introduction, in the case of target-culture taboos or governmental request, such as the exposure of bones being prohibited in Chinese. Therefore, the sections pertaining to translation are Card Name, Type Line, and Text Box (which is in turn separated in Card Abilities, Reminder Text, and Flavor Text). In order to maintain in-game consistency, the text translated within Type Lines, Card Abilities, and Reminder Text is usually more formulaic and less prone to being changed throughout editions, even though the translation options chosen in its Fourth Edition may not be the best options currently available. Card Names usually accept greater creative endeavors, with careful checking of the 20,000 previously released cards in order to prevent card name repetition for different cards. This has recently proven to 65 be a specific translational conundrum, as essayed by the author in one expansion which does not involve the studied corpus11.

4.2.3 Internal narrative structure (reflection of plot onto the cards), dramatic plot (prologue, exposition, climax, conclusion, epilogue – reflected in the cards) Any reference to the game plot story is implied in each card by what it performs in play, but it is only clearly stated in its flavor text. In the game official website, short stories are also released weekly, in order to provide further plot insight which is clearly indicated by utilizing card illustrations to illustrate the short stories themselves. Until the New Phyrexia12 block, digital comic books were released in the official website, but they are on a hiatus currently lasting over 4 years. One more plot endeavor is the release of novels for specifically flavorful narrative blocks, such as Return to Ravnica. Three novels have been released for Return to Ravnica, each with the cover illustration of its respective expansion booster pack. The online comic books were translated to Portuguese, and later to Spanish in the last 2 editions; the novels have been released only in English. However, this pertains mostly to narrative blocks of 3 expansions each, and Core Sets have less plot information due to their function as transitioning between the previous block and the next. Since players are portrayed as planeswalkers who travel through different planes of existence, the idea may be that of walking from the previous plane (where the last story unraveled) to the other, through the Æther that flows between planes of existence, and on its way to the next plane. In the case of Magic 2013, the Innistrad block was being wrapped up: based on European and particularly Germanic Grimm-like horror, this plane had been one of the most popular among players, having broken sales records. M13 was also preparing the field for the Return to Ravnica block – being Ravnica a city- world governed by ten guilds, each with their own agenda, which had already been a narrative block years before. Both planes can be glimpsed through some of the M13 cards, as if flowing from one plane into the other.

11 Meg’s Gathered Translations – Tumblr. Available in Last access on Dec 16th 2013. 12 Narrative block released in 2011, returns to the plane of Mirrodin (2003) and develops the plot of the invasion of Mirrodin by the Phyrexians, a race that infects the plane’s inhabitants. 66

This strategy of using a Core Set to stabilize the narratives between blocks has been used for the last 4 years, starting with Magic 2010; before that, the main narrative concentrations were called Editions, but they were less regular in release as the game developed, especially during its first 4 years. It is very important to note that in play, plot understanding is not fundamental or necessary, being ignored by several players. Plot is of particular interest to the Vorthos, a player archetype who enjoys the story more than any other aspect of the game, and plays in accordance with it (Cavotta, 2005). In the view of translation, the flavor text is the place where most creativity can flow, and a more local transcreation could be performed, but it seems that this chance is not taken advantage of. Most translated flavor seems to be produced to be adequate, but could be produced to be acceptable at least in this section, in which the narrative is more important than the game rules.

4.2.4 Author comments (Framing Information) on rules description As mentioned earlier, the only presence of an analogue to author footnotes is in what the game refers to as reminder text. This is most commonly used for new mechanics; for example, Exalted is the new mechanics introduced in Magic 2013, and its explanatory text can be seen in Figure 12. There is no space left in the card for translational footnotes, unfortunately.

4.3 Microstructure Analysis The macrostructure analysis has given further insight on the adequate translation strategy likely employed to localize this product. In this section, cards from the corpus will illustrate some punctual corpus findings, and will be given a closer look.

67

4.3.1 Lexical choices

Figure 14: Word cloud with the 50 most frequent words in the analyzed corpus and their respective frequencies. Grammatical words of Brazilian Portuguese language have been ignored. Created at tagcrowd.com

The most frequent lexical choices in this corpus can be seen in Figure 14 above. They pertain to 4 essential semantic groups, as seen on Table 7 below: dueling, magic, metagame, and its collecting element. Words with no particular category (mesmo, qualquer) will be considered as grammatical words that have not been properly ignored by the system.

68

Table 7: Semantic sorting of the 50 most frequent words in the studied corpus, as defined by the author Dueling Magic Metagame Collecting Alvo Conjurar Card(s) Comum Artefato Feitiço Consulte Incomum Atacar Grimório Custom Raro Batalha Instantânea Deck Campo Mágica(s) Etapa Cemitério Mana Fase Combate Planeswalker Jogador(es) Controla Vida Jogar Criatura(s) Voar Jogo Dano Magic Efeito Mão Habilidade Página Oponente Permanente Terreno(s) Tipo Turno

In turn, the most frequent words in the corpus do not properly reflect the sophisticated vocabulary that is pertaining to Magic as a Research and Development prerogative, at least in its development language (Rosewater, 2012). According to Rosewater, current lead designer of Magic, it is part of their minor objectives in game design to expand the players’ vocabulary and introduce them to interesting, uncommon, and rare words of English language. By analyzing the corpus, particularly on what I call its creative sections (namely, Card Name and Flavor Text), it is possible to assume that this is also followed in Portuguese. Some examples of uncommon words used in Portuguese-localized card names from the corpus can be seen on Table 8 below, accompanied by their also uncommon lexis in English. The sophisticated vocabulary may be also another way to reinforce the element of fantasy in which the stories of Magic are developed. 69

Table 8: 10 sample card names from the studied corpus which exemplify the sophisticated vocabulary used in both languages, implying the use of such vocabulary as one localization practice. English Portuguese Spiked Baloth Baloth Espiculado Elderscale Wurm Vorme Vetusto Welkin Tern Carrago Celeste Roaring Primadox Primadox Rugidor Hellion Crucible Crisol de Avernais Void Stalker Espreitador do Vácuo Unsummon Esconjurar Goblin Arsonist Goblin Pirômano Rootbound Crag Penhasco do Raizame Duress Coagir

4.3.2 Dominant grammatical patterns Being a game in which most of the text gives instructions on how the rules can be enacted in play, most verbs are conjugated in the imperative tense when obligatory rules, and in conditional forms when it is the choice of players to be activated or not. This can be observed clearly in the Text Box, particularly in Card Abilities and Reminder Text, as explained in the Introduction (see examples of mechanics keywords and their respective reminder text from the corpus in both languages in Table 9 below).

70

Table 9: 10 sample mechanics keywords from the studied corpus, accompanied by their reminder text. English Portuguese Trample (If this creature would assign Atropelar (Se esta criatura fosse atribuir enough damage to its blockers to dano suficiente para destruir seus destroy them, you may have it assign the bloqueadores, você pode fazer com que rest of its damage to defending player or ela atribua o resto de seu dano ao planeswalker.) planeswalker ou jogador defensor.) First strike (This creature deals combat Iniciativa (Esta criatura causa dano de damage before creatures without first combate antes das criaturas sem strike.) iniciativa.) Reach (This creature can block creatures Alcance (Esta criatura pode bloquear with flying.) criaturas com voar.) Vigilance (Attacking doesn't cause this Vigilância (Esta criatura não é virada creature to tap.) para atacar.) Exalted (Whenever a creature you Exaltado (Toda vez que uma criatura control attacks alone, that creature gets que você controla ataca sozinha, ela +1/+1 until end of turn.) recebe +1/+1 até o final do turno.) Regenerate (The next time this creature Regenere (Na próxima vez em que esta would be destroyed this turn, it isn't. criatura seria destruída neste turno, ela Instead tap it, remove all damage from não será. Em vez disso, vire-a, remova it, and remove it from combat.) todo o dano dela e remova-a do combate.) Lifelink (Damage dealt by this creature Vínculo com a vida (O dano causado also causes you to gain that much life.) por esta criatura faz com que você ganhe uma quantidade equivalente de pontos de vida.) Defender (This creature can't attack.) Defensor (Esta criatura não pode atacar.) Equip (Attach to target creature you Equipar (Anexe à criatura alvo que você control. Equip only as a sorcery.) controla. Equipe somente quando puder conjurar um feitiço.) Flying13 (This creature can't be blocked Voar (Esta criatura só pode ser except by creatures with flying or bloqueada por criaturas com a reach.) habilidade de voar ou alcance.)

In turn, mechanics keywords present less grammatical consistency. Mechanics are keywords used to summarize card abilities which are

13 Flying did not present reminder text in Magic 2013; source of reminder text from Tenth Edition: Gatherer.wizards.com 71 frequently used in play, and that enact specific effects. In every expansion or set there is at least one new mechanics and a few returning ones (being Exalted the case for M13) and, at least in the first few releases of the mechanics, reminder text is added onto the card text boxes in order to clarify what the keyword means and how it should be played. Although that is dependent on the amount of text in the box, flavor text is not provided for cards which need more room for rules clarification. It is important to note that this is also done in the developmental language and therefore does not classify as translator’s footnotes or Framing Information (Gile, 2009), although they may seem to be so. Additionally, when more text is needed, font sizes can be reduced in order to fit more text – again, a tactic that is also performed in the developmental language when necessary. Font sizes that pertain only to translation will be granted further comment on Section 4.5, Technical Constraints. With the examples above, it is possible to clearly observe the grammatical inconsistency in the translation of keyword abilities. Among these examples, three ability types are present: static abilities (abilities that are always functioning, as long as a card with the ability is in play), triggered abilities (abilities that are only functioning depending on the occurrence of a specific effect in the game) and activated abilities (abilities which require certain player actions to function, such as paying mana, sacrificing creatures, discarding or buying cards, and so on). The activated abilities in Table 10 are Regenerate and Equip; Exalted is a triggered ability, while the remaining others are static abilities. Since they serve different functions in play, they should receive different grammatical treatment: triggered abilities are translated as adjectives because they become, say, exalted, when and only when something happens in play – similarly to what is done in the development language. Activated ability keywords are translated as verbs because they are also verbs in the development language – but also because they are actions to be performed. In the case of static keyword abilities, they are not actions to be performed; they are innate characteristics of each card and, therefore, should be translated as nouns only. This grammatical class would imply that this is an ability the card intrinsically has and, depending on the effect of certain cards, it can be gained or lost. However, there is grammatical ambiguity in the development language as regards to that, because the word flying for example can be used either as a verb or as an adjective depending on the context. This may be a possible reason for the 72 translation of some static ability keywords to have been conjugated as verbs, particularly abilities such as Flying and Trample which have been in use since the very beginning of the game (Alpha Edition), before it started being localized to Brazilian Portuguese in its Fourth Edition. It is possible to imply from the data in Table 10 that more recent keywords have been localized according to the observations above, as the remaining static abilities (Iniciativa, Alcance, Vigilância, and Vínculo com a vida) are translated as nouns. This reflects their intrinsic feature clearly, leaving room for these abilities to be gained or lost depending on certain card effects, as can be seen in Figure 15 below.

Figure 15: An Instant-type card which grants a static ability (First Strike) temporarily, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size).

4.3.3 Speech reproduction (direct versus indirect) Flavor text is the card section in which the strongest narrative elements can be found; flavor-drawn players tend to rely on these small segments of narrative to complete the puzzle of what story the expansions are telling. In this section, character quotes are often present; they feature both big and small characters of the Magic multiverse, from Planeswalkers to Legendary Creatures, as well as minor characters who appear as Creature 73 card names, but no specific reference to their name. An example of this third occurrence can be seen in Figure 16 below; the War Priest of Thune is another creature card with no reference to the name Idrus, who is quoted in Faith’s Reward, using direct speech in both languages.

Figure 16: An Instant-type card which flavor text references a minor character from another card named War Priest of Thune, and its localization to Portuguese (Source: Gatherer.wizards.com – actual size).

4.3.4 Narrative, perspective, and point of view Also on the flavor text section, flavor text comes not only in the form of character quotes, but also that of narration. This narration can be seen in the forms of omniscient narration, camera narration, from the point of view of one or more characters, seldom as a fictional citation of an in-multiverse publication, and more rarely still a real-life or even literary reference (see Figure 17 below). This multiplicity of perspectives reinforces the multiplicity of characters and worlds involved in Magic, and that is usually conveyed with similar accuracy in flavor translation. The translation of this culture-specific flavor text maintains the real-world or literary reference even though the translation is still performed as adequate. This is one particular point in which a shift to appropriate translation could make the adaptation of the real-world reference stronger, easier to recognize and 74 relate, and furthermost more special for player experience.

Figure 17: An Instant-type card which flavor text makes a real-world literary reference to Hamlet, by William Shakespeare, and its localization to Portuguese (Source: Gatherer.wizards.com – actual size).

4.3.5 Modality (passive versus active, uncertainty, ambiguity) For the localization of a rules-based game such as trading cards, the wording is crafted in the sense that ambiguity is reduced to a minimum. However, it is possible for doubts to emerge during play, particularly when two or more cards are in action simultaneously (in what the Basic Rulebook calls ‘the stack’). Such doubts can be solved by a judge during sanctioned events, among other players, or through consulting the Discussion section of the respective cards in Gatherer, the Magic official database. Translation is performed with similar care; however, in past and later editions, there have been translation errors which caused the judges to be warned of the functioning of the card as described in the development language, but not as translated due to improper translation which caused change in rules.14

14 Fornazari, M. R. (2013) Meg's Gathered Translations - Tumblr. Available on accessed on 01/14/2014. 75

4.3.6 Language levels (archaic, popular, or dialect; game jargon) Although the game is marketed for 13-years-old or above, the language level used in Magic The Gathering is not facilitated for the reading capacity of preteen audiences. That may be due to the fact that a player age average for the general gamer is of approximately 30 (Bernal- Merino, 2006), but also in order to reinforce the fantasy environments on which the game is narratively based. An aforementioned article by Mark Rosewater (2012) reinforces their interest in creating a game that also serves to expand the vocabulary of players and strengthen game immersion through it. In translation, the same trend can be observed in the use of uncommon and rare wording in Portuguese as well (see Table 8 for examples), which implies that this could be part of the localization briefing. Additionally, the language level of the game has a strong element in its jargon, as it functions as metalanguage about the game itself. In the case of Magic, jargon would include its card elements, match elements, and game mechanic keywords. Card element jargon can be seen on Figure 16, and mechanics keywords on Table 10, followed by their respective micro analysis; the jargon pertaining to the segments of a match can be exemplified in Figure 18 below, which lists the phases and steps of each turn in a match as described in the Basic Rulebook (2012). Translation choices also point to a more adequate translation trend, especially because of the normative character of these elements in play. This also implies the use of Translation Memory systems, due to the similarity between segments in both lists.

76

Figure 18: Parts and steps of a turn during a match, extracted from the M13 Basic Rulebook, and its localization to Portuguese (Source: wizards.com).

4.4 Systemic Context Analysis To simply find the typical occurrences and possible practices of a translator is not enough to understand why such decisions have been made (Hermans, 1996). One needs to go deeper. According to Bernal-Merino (2006), statistical data shows that the age average of gamers is of 33 years old, being 38% female. It is likely that, due to the DCI registration (Duelists' Convocational International, an organization created by Wizards of the Coast to organize competitive player tournaments and achievements), the company has access to statistical data that specifically pertains to Magic15. Additionally, the female Magic player percentage is expected to be lower than in the average gaming industry (estimated below 10%, based on the author's personal experience as a casual player). This may influence the existence of complex vocabulary and rules system, although several players have been introduced to the game as

15 Unfortunately, the marketing department in Wizards of the Coast (WotC) cannot disclose such data, according to personal communication (e-mail) with WotC brand manager Jennifer Meyen. 77 children by a parent or relative at about 8 years of age, or as early as they gained reading proficiency in order to understand the card text on their own16. The microstructural analysis in the previous section has provided more data which should be cross-referenced with the macrostructural data, in order to consider the broad systemic context (Lambert & Van Gorp, 1985).

4.4.1 Macro versus microstructure; text versus theory (norms or models) Both micro and macro structure data have indicated consistency in producing adequate translations in the Trading Card Game, due to its normative functioning in play. Additionally, in both analyses, it was possible to observe that there seems to be more transcreative space in sections which do not reflect such normative function (what I have called creative sections: Card Name and Flavor Text); however, although such transcreation room exists, the translated product is still adequate when it could be acceptable in these sections. Cultural adaptations and changes could be of advantage when the translated result seems to be distant from the target culture. In order to illustrate this, Figure 19 below presents a card with flavor text that makes a reference to the American expression “it’s like Christmas in July”; Christmas may have been replaced by birthday because Christianity references would not fit the context of the game, nor our real- world forms of calendar, although the expression sounds like the original similarly enough. However, this is an idiom which could have been translated differently, as the adequate translation does not reflect culturally the joy of Christmas in July, or Christmas come early. Transcreation to create a flavor text that is acceptable in the Brazilian cultural system could be of use in such moments.

16 Based on the author’s participation in a MTG player community in Southern Brazil, from which such examples of early players have been drawn. 78

Figure 19: A Sorcery-type card in which the flavor text makes a real-world cultural reference to the American expression it’s like Christmas in July, and its adequate localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size).

4.4.2 Intertextual relations Although the intertextual relations in the Magic multiverse are also related to flavor text, other forms of intertext are not currently translated. As mentioned earlier, the short stories in the official Magic website are not translated, nor the novels released about the narrative blocks (such as the three Return to Ravnica block novels by Doug Beyer, released with each of the three expansions between 2012 and 2013). There are no novels about the Core Sets, and there have been no novels released about the Innistrad block, the narrative expansions previous to the Magic 2013 Core Set. Also, the digital comic books have been discontinued in 2009. However, other intertextual relations could be drawn among cards in the same set. The flavor text and/or card names in several cards present quotes or references to planeswalkers or legendary creatures, such as Crusader of Odric (presented in Figure 20 below) which refers to Odric, Master Tactician (see Figure 1). Such intertextual relation is not innocent as the flavor also indicates a fruitful combination of both cards in play. 79

Translation is performed accordingly, maintaining intertextual references as such.

Figure 20: A Creature-type card in which both card name and flavor text make an in-game reference to another card in the corpus named Odric, Master Tactician, and its localization to Brazilian Portuguese (Source: Gatherer.wizards.com – actual size).

4.5 Technical constraints By observing the card images in the multimodal corpus, it is possible to note font reduction in 41% of the Portuguese-localized cards (103 out of 249). In Chandler and Deming (2012) there is mention to 20-30% text being larger in translated text, therefore taking up more space; however, it is not stated clearly if this is due to explicitation choices on behalf of the translator or due to language or genre specificities. The solution seems to have been to reduce font size in up to 30% in order to fit the same Message in Portuguese. This is also common practice for texts localized to French and German, resulting in text approximately 30% longer when translated from English (Esselink, 2000). It is possible to assume that this is an automated response to the larger amount of text, as the translated card text may be added onto a card template in order to fit into the previously 80 allocated spaces. The template can adapt font sizes to the amount of added text, as can be seen in informal card-creating software such as Magic Set Editor (Twan Van Laarhoven, 2011). This kind of software is used often to create card parodies of no playing value and custom playable tokens of no selling value. Chandler and Deming (2012) mention the use of icons as a widespread technique used by game developers in order to simplify certain meanings within a game; in the case of Magic, certain symbols have been created and are now in-game canon, which are the mana and card-tapping symbols (wubgrt). Within the game, they have their own meaning, and very rarely there is need for the card rules to literally state for a Mountain card, "T: Add R to your mana pool", having only R instead, with the same meaning (see Figure 2). These icons serve as nonverbal information which not only simplifies meanings in the game, but also simplifies the translation of such meanings. This is also true for card copyright information, collector number, and expansion symbol (see Figure 13) which are also not translated elements which carry their meaning by using numbers and symbols.

4.6 Translation Policies From the results observed and presented in this chapter, based on the analyzed corpus, it is possible to observe the Magic 2013 corpus in both English and Portuguese very closely and derive the following possible constraining translation policies: 1. The translation seems to be produced to create a product that is acceptable as text in Brazilian Portuguese language; 2. Due to the flavor narrative of fictional fantasy universes, the translation may be discouraged to produce real-world culturally- specific references, unless the development language has already done so; 3. More recent mechanics keywords for static abilities have been translated as nouns, sustaining their meaning as an ability that is innate, received, or lost; activated abilities are translated as verbs as they are performed during play. 4. Text that has been translated in previous editions is hardly ever changed, in order to maintain rules consistency; 81

5. Font sizes may not be an issue for translators to actively decide on, as they may be automatically adapted or adapted by visual design engineers; 6. Game text (Card Type, Rules Text, Reminder Text) is more formulaic and algorithmical, whereas creative text (Card Name and Flavor Text) is not. These policies could be extrapolated as representatives of its genre; however, a suggestion for future TCG scholars would be to analyze other card games in order to find consistencies (and better yet, inconsistencies) with Magic The Gathering.

82

83

5 CONCLUSION

84

This chapter will present a recap of the findings derived from the study performed, while delineating its limitations and indicating other research paths that could be followed in future studies, or by other scholars.

5.1 Summary of the main points From the results presented in the previous chapter, it is possible to draw the following conclusions on the observation of the Magic 2013 corpus constructed for this study: 1. At several points of the corpus, the acceptable translation produced can make the text seem truncated and not applicable to its tenor of being written to be played. 2. When the development language has real-world North-American cultural references, they are still translated to be acceptable when they could be translated to be adequate to the Brazilian cultural point of view; when they are not, they sound like a caricature of American culture in which sometimes creative flavor can be lost in humor or meaninglessness (as seen in Section 4.4.2). 3. Mechanics keywords for static abilities should be translated as nouns; activated and triggered abilities as verbs. 4. Mechanics keywords that do not follow the aforementioned grammatical consistency to rules and have been translated in previous editions should be changed, in order to achieve grammatical consistency. 5. Font sizes may not be decided by translators, but should be considered as font reduction may be avoided by the use of translation strategies. 6. Creative text (Card Name and Flavor Text) should be translated in order to create culturally acceptable translated sections, instead of linguistically adequate sections like the game text in Trading Card Games. These conclusions could be extrapolated to other Trading Card Games; however, the aforementioned suggestion for future TCG to be analyzed is still of importance in order to check the reproducibility of such conclusions.

5.2 Limitations of the study Due to time constraints, only one Core Set of cards has been analyzed in this study; this has resulted in a corpus that seems to be too 85 small to be representative. However, one must consider the small amount of text in the analyzed media to recognize its relative representativeness. Although Trading Card Games are a multimodal medium, the corpus created only comprised the textual elements of each card and Rulebook. Until the final writing stage of this study, it was impossible to upload it to the online corpus COPA-TRAD due to technical problems with the website system administration. Additionally, only the textual elements could not have card images added as a part of the corpus itself, due to the complexity of such task within the website system. It is expected that at least the textual elements of the present corpus will be made available at COPA-TRAD.

5.3 Suggestions for future research In relation to the present relatively small corpus size (see Chapter 4, Table 5), it could be interesting to perform a study which would comprise not only one Core Set, but also the three narrative Expansions that follow it during a year, with that year’s Rulebook as metatext. This would allow not only for a more representative corpus, but also for other levels of analysis, such as the construction of a narrative through cards which carry approximately 10 words of narrative flavor text per card, for example. Additionally, future studies using Magic The Gathering as research object as well may later enable its becoming a dynamic corpus, adding both previous and future editions of Magic to the corpus, or even other Trading Card Games. This would be of academic interest as it would make available a longitudinal analysis of the genre in the sense of its linguistic and translational evolution. For example, in previous editions the term planeswalker had been translated to planinauta, only to appear un- translated again in more recent editions; this may be due to copyright issues. This supposition will remain unchecked as the scope of this study does not comprise the editions in which the term planeswalker had been translated.17 Each card image could be added onto the corpus in order to maintain visual aspects, design, text positioning, and other details in order to make the text as authentic as possible, as advised by Leech (1992). In addition,

17 These translated cards have been printed in previous editions, such as Planeshift (2001). Although the official MTG database does not have all card images in Portuguese, unofficial databases such as magiccards.info make such translated versions available. 86 the image files may be of further interest as the card illustrations can be used for later multimodal studies, as performed briefly by Fornazari (2013). However, there is room for further multimodal studies in relation to, for example, gender and ethnic representation in card illustrations. Finally, it would be of great interest for scholars specialized in other languages to add other localizations of Magic The Gathering to the corpus, transforming it into a multilingual corpus which could enable similar studies to be performed with any of the other 9 languages to which the game is localized. Also, comparative translation studies could be performed among, say, Portuguese, Spanish, Italian, and French in order to find best practices for Romance language localization in Trading Card Games. In relation to other aspects of localization and not only of corpus- based studies, it is suggested for future studies to add the voice of a number of translators and localizers previously or presently involved with such type of product. Semi-structured interviews could be performed with these individuals in order to create a localizator profile and check if the translation practices observed by the corpus are in fact part of their everyday work. Additionally, reception studies can be performed with Trading Card Game players in order to assess if the localization choices are actually used or not, and the reasons why it is or is not being used. This can be performed through semi-structured interviews and also through the recording of matches, in order to analyze the translation options (or lack thereof) used during the declarative discourse commonly used during a match. 87

References

Baker, M. (Ed.). (2009). The Routledge Encyclopedia of Translation Studies. New York: Routledge. Baker, M. (1995). Corpora in Translation Studies: An Overview and Suggestions for Future Research. Target, 7(2), 223-243. Berman, A. (2004). La traduction comme épreuve de l’étranger [translated by L. Venuti as Translation and the trials of the foreign]. In: L. Venuti (Ed.). (2004). The Translation Studies Reader. New York: Routledge. pp. 240-253. Bernal-Merino, M. A. (2006). On the Translation of Video Games. In The Journal of Specialized Translation 6, pp. 22-36. Retrieved from http://www.jostrans.org/issue06/art_bernal.php Blizzard Entertainment. (2011). World of Warcraft lançado no Brasil. Retrieved from http://us.battle.net/wow/pt/blog/4054272/World_of_Warcraft_lan%C 3%A7ado_no_Brasil-06_12_2011 Bulfinch, T. (2002). O Livro de Ouro da Mitologia [translated by David Jardim Júnior]. Rio de Janeiro: Ediouro. Cavotta, M. (2005). Snack Time With Vorthos. Daily MTG Magazine. Retrieved from https://www.wizards.com/Magic/Magazine/Article.aspx?x=mtgcom/ daily/mc18 Chandler, H. M. & Deming, S. O. (2012). The Game Localization Handbook. Jones & Bartlett Publishers. Coletti, B. (2013). COPA-MUM - A importância do Domínio da Linguagem Especializada na Localização de Roleplaying Games. In Vidal, C.D., Elias, I.M., Heberle, V.M. Pesquisas em Games: Ideias, Projetos e Trabalhos (pp. 57-67). Florianópolis: UFSC/LLE/CCE. Cronin, M. (1998). The cracked looking glass of servants: Translation and minority in a global age. The Translator 4(2), pp. 145-162. Cunningham, M. A. (2013). Inside Gaming: Interview with former Square Enix translator Tom Slattery - RPGamer. Retrieved from http://www.rpgamer.com/features/insidegaming/tslatteryint.html Donovan, T. (2012). 10 Tips: Localization. Gamasutra Featured Article. Retrieved from http://www.gamasutra.com/view/feature/175367/10_tips_localizatio n.php?page=1 88

Eco, U. (2003). A Componential Analysis of the Architectural Sign /Column/. Semiotica 5(2). pp. 97-117. Edwards, K. (2011). Culturalization: the Geopolitical and Cultural Dimension of Game Content. Trans 15 pp. 19-28. Esselink, B. (2000). A Practical Guide to Localization. John Benjamins Publishing Company. Fernandes, L. P. (2004). Brazilian practices of translating names in children's fantasy literature: a corpus-based study. PhD dissertation, Programa de Pós-Graduação em Inglês, Universidade Federal de Santa Catarina. Florianópolis. Fornazari, M.R. (2013) Anúncio de Pesquisa: "Magic the Gathering" Localizado do Inglês Para o Português. In Vidal, C.D., Elias, I.M., Heberle, V.M. Pesquisas em Games: Ideias, Projetos e Trabalhos (pp. 286-291). Florianópolis: UFSC/LLE/CCE. Fornazari, M.R. (2013). Magic the Gathering sob a Ótica da Gramática Visual. In-traduções 5. pp. 13-28. Fornazari, M. R. (2013) Meg's Gathered Translations - Tumblr. Retrieved from meghan1337.tumblr.com Garfield, R.C. (1997). Trading Card Game Method of Play. US Patent #5,662,332. Gile, D. (2009). Fidelity in Interpretation and Translation. Basic Concepts and Models for Interpreter and Translator Training, Volume 8, pp. 52-78. Amsterdam: John Benjamins Publishing. Grande, P. (2013) Vocabulario en juegos: Magic Laberinto del Dragón - Un Arácnido, Una Camiseta. Retrieved from http://www.unaracnidounacamiseta.com/2013/04/25/vocabulario-en- juegos-magic-laberinto-del-dragon/ Harper, E. (2007). The9 changes WoW in China to appease censors. WoW Insider. Retrieved from http://wow.joystiq.com/2007/07/03/the9- changes-wow-in-china-to-appease-censors/ Hartley, T. (2009). Technology and Translation. In Munday, J. (Ed.), The Routledge Companion to Translation Studies. New York: Routledge. pp.106-127. Hermans, T. (1996). Norms and the determination of translation: A theoretical framework. pp. 25-51. Holmes, J. (1972) The Name and Nature of Translation Studies. In: Venuti, L. (2000). (Ed.) The Translation Studies Reader. New York: Routledge. pp. 172-185. 89

Honeywood, R. (2011). Best Practices in Game Localization. International Game Developers Association. Retrieved from http://englobe.com/wp-content/uploads/2012/05/Best-Practices-for- Game-Localization-v21.pdf Lambert, J. and Van Gorp, H. (1985). Sobre a descrição de traduções. [Translated by Torres, M. H. C. & Fernandes, L. P.] In Guerini, A., Torres, M. H. C. and Costa, W. (Org.) Literatura & Tradução: textos selecionados de José Lambert. Rio de Janeiro: 7Letras, pp. 197-213. Lee, A. (2013). The state of games localisation. Develop 137. Retrieved from http://www.develop-online.net/features/1869/The-state-of- games-localisation Leech, G. (1992). Corpora and theories of linguistic performance. In Svartvik, D. (Ed.) Directions in Corpus Linguistics: Proceedings of Nobel Symposium. Berlin: Mouton de Greyter. pp.105-122. MacInnes, I. (2006). Property rights, legal issues, and business models in virtual world communities. Electron Commerce Res 6. pp. 39-56. Mangiron, C. & O'Hagan, M. (2006). Game Localisation: Unleashing Imagination with 'Restricted' Translation. The Journal of Specialized Translation (6), pp. 10-21. Retrieved from http://www.jostrans.org/issue06/art_ohagan.php Mangiron, C. & O'Hagan, M. (2013). Game Localization: Translating for the Global Digital Entertainment Industry. London: John Benjamins. Maragos, N. (2009). The Process - ATLUS Production Diaries. Retrieved from www.atlus.com/forum/showthread.php?t=4016 Mason, I. (2001). Translator Behaviour and Language Usage: Some Constraints on Contrastive Studies. Hermes, Journal of Linguistics 26, pp. 65-80. McCurley, M. (2011) The Lawbringer: Where localization meets legal reasoning - WoW Insider. Retrieved from http://i.wow.joystiq.com/2011/06/10/the-lawbringer-where- localization-meets-legal-reasoning/ Munday, J. (Ed.) (2009). The Routledge Companion to Translation Studies. New York: Routledge. Nasr, S. (2006). Dez itens curiosos do menu internacional do McDonald’s – Como Tudo Funciona UOL. Retrieved from http://lazer.hsw.uol.com.br/10-itens-menu-internacional- mcdonalds.htm 90

Newmark, P. (2004). Non-literary in the Light of Literary Translation. The Journal of Specialised Translation 1, pp. 8-13. Notepad++ Team (2013). Notepad++ [software] Olohan, M. (2004). Introducing Corpus-based Translation Studies. New York/London: Routledge. Palumbo, G. (2009). Key Terms in Translation Studies. London: Continuum. pp.71-72. Pym, A. (1996) ‘Venuti’s visibility’ (Review of The Translator’s Invisibility), Target 8.1: 165–77. Rocha, N.F.F. (2013). Olha que coisa mais linda: As Traduções da Canção 'Garota de Ipanema' em Inglês, Alemão, Francês e Italiano sob a Ótica do Sistema de Transitividade. Masters Thesis, Programa de Pós-Graduação em Estudos da Tradução, Universidade Federal de Santa Catarina, Florianópolis. Rosewater, M. Mark Rosewater Archive - Daily MTG Magazine. Retrieved from https://www.wizards.com/Magic/Magazine/Archive.aspx?author=Ma rk%20Rosewater Schell, J. (2008). The Experience Rises Out of a Game. In The art of game design: a book of lenses. Chapter 3. Burlington, MA: Morgan Kaufmann. pp. 23-38. Scott, M. Wordsmith 3.0 [software] Shakespeare, W. (1603). Hamlet, Prince of Denmark. Skoog, K. E. (2012). When to Forgo the Culturalization of Video Games: Contextualizing Globalization within the Mobile Marketplace - Gamasutra (Featured Post). Retrieved from http://www.gamasutra.com/blogs/KarinESkoog/20120702/173371/ When_to_Forgo_the_Culturalization_of_Video_Games_Contextuali zing_Globalization_within_the_Mobile_Marketplace.php Sternberg, R.J. and Lubart, T.I. (1999). The Concept of Creativity: Prospects and Paradigms. In Sternberg, R.J. (Ed.). Handbook of Creativity. Cambridge: Cambridge University Press, pp. 3-15. Thunes, M. (1998). Classifying Translational Correspondences. In Johansson, S. and Oksefjell, S. (1998) Corpora and Cross-Linguistic Research: Theory, Method and Case Studies. Amsterdam: Rodopi. pp. 25-50. 91

Toury, G. (1995). The Nature and Role of Norms in Translation. In Venuti, L. (2012). (Ed.) The Translation Studies Reader. New York: Routledge. pp. 198-212. Trammell, A. (2010) Magic: The Gathering in material and virtual space: An ethnographic approach toward understanding players who dislike online play. Michigan State University Meaningful Play proceedings. Retrieved from http://meaningfulplay.msu.edu/proceedings2010/mp2010_paper_42. pdf Van Laarhoven, T. (2011). Magic Set Editor [software] Venuti, L. (2012). The Translation Studies Reader. New York: Routledge. Vidal, C.D. and Elias, I.M. (2012). When Brazil enters the Koprulu sector: the Starcraft II localized into Brazilian Portuguese. Proceedings of XI SBGames 2012. pp. 114-124. Vidal, C.D. (2011). Multimodalidade e traduções funcionais para investigar a aquisição de segunda língua em gamers. Doctoral dissertation. Pós- Graduação em Estudos da Tradução, Universidade Federal de Santa Catarina. Vidal, C.D. (2013). A Experiência da Localização de Games da Blizzard ao Português do Brasil: um Estudo de Caso. Post-Doctoral study. Pós- Graduação em Estudos da Tradução, Universidade Federal de Santa Catarina. Williams, J. & Chesterman, A. (2002). The Map - A Beginner's Guide to Doing Research in Translation Studies. Manchester, United Kingdom: St. Jerome Publishing. Wizards of the Coast Press Release, May 9 2003 - Daily MTG Magazine. Retrieved from http://www.wizards.com/Magic/TCG/Article.aspx?x=magic/magicon line/news050903 Wizards of the Coast. (2012). M13 Basic Rulebook. Retrieved from http://media.wizards.com/images/magic/resources/rules/EN_MTGM 13_Rulebook.pdf Wizards of the Coast. Gatherer - The Magic Card Database. Retrieved from gatherer.wizards.com Wizards of the Coast. How to Play - Magic the Gathering. Retrieved from http://www.wizards.com/Magic/TCG/NewtoMagic.aspx?x=mtg/tcg/ newtomagic/learntoplay 92

Whited, L.A. (2004). The Ivory Tower and Harry Potter: Perspectives on a Literary Phenomenon. Columbia: University of Missouri Press.